EA042157B1 - NON-INVASIVE DETERMINATION OF FETUS METHYLOMA OR TUMORS ON PLASMA - Google Patents

Description

Cross-references to related application

This application is a PCT application claiming priority over U.S. Provisional Application No. 61/830571 entitled: Tumor Detection In Plasma Using Methylation Status And Copy Number filed 3 June 2013; and U.S. Patent Application No. 13/842209 entitled: Non-Invasive Determination Of Methylome Of Fetus Or Tumor From Plasma, filed March 15, 2013, which is a non-tentative application and claims priority US Provisional Application No. 61/703512 entitled: Method Of Determining The Whole Genome DNA Methylation Status Of The Placenta By Massively Parallel Sequencing Of Maternal Plasma filed September 20, 2012, which are hereby incorporated by reference in their entirety in all respects.

Technical field

The invention disclosed herein relates generally to the determination of the methylation pattern (methylome) of DNA, and more specifically to the analysis of a biological sample (e.g. plasma) that contains a mixture of DNA from different genomes (e.g. fetal and maternal, or tumor and normal cells) , to determine the methylation pattern (methylome) of the minor genome. Also described are options for using the established methylome.

State of the art

The development of the embryo and fetus is a complex process involving many highly coordinated genetic and epigenetic events. The development of cancer is also a complex process involving a number of genetic and epigenetic steps. Anomalies in the epigenetic control of developmental processes are associated with infertility, spontaneous abortions, intrauterine growth anomalies, and postnatal consequences. DNA methylation is one of the most intensively studied epigenetic mechanisms. DNA methylation mainly occurs in the context of adding a methyl group to the 5' carbon of cytosine residues in CpG dinucleotides. Cytosine methylation adds another layer of control over gene transcription and DNA function. For example, hypermethylation of gene promoters rich in CpG dinucleotides, called CpG islands, is generally associated with downregulation of gene function.

Despite the important role of epigenetic mechanisms in mediating developmental processes, human embryonic and fetal tissues are not readily available for analysis (similarly, tumors may not be). Studies of dynamic changes in such epigenetic processes in normal and in the presence of diseases during the prenatal period in humans are practically impossible. One of the main directions of such research was represented by extraembryonic tissues, in particular, the placenta, which can be obtained during prenatal diagnostic procedures or after the birth of a child. However, these tissues require invasive procedures.

The DNA methylation profile of the human placenta has been of interest to scientists for many decades. The human placenta exhibits many peculiar physiological features involving DNA methylation. Globally, placental tissues are hypomethylated compared to most somatic tissues. At the gene level, the methylation status of individual genomic loci is a specific signature of placental tissues. Both global and locus-specific methylation profiles show gestational age-dependent changes. Imprinting genes, namely genes whose expression depends on the parental origin of alleles, perform key functions in the placenta. The placenta has been described as premalignant and hypermethylation of several tumor suppressor genes has been observed.

Studies of the DNA methylation profile of placental tissues have provided valuable information for understanding the pathophysiology of pregnancy-associated or developmental diseases such as pre-eclampsia and intrauterine growth retardation. Genomic imprinting disorders are associated with developmental disorders such as Prader-Willi syndrome and Angelman syndrome. Altered profiles of genomic imprinting and global DNA methylation in placental and fetal tissues have been observed in assisted reproductive technology pregnancies (H Hiura et al. 2012 Hum Reprod; 27: 2541-2548). A number of environmental factors such as maternal smoking (KE Haworth et al. 2013 Epigenomics; 5: 37-49), maternal nutrition (X Jiang et al. 2012 FASEB J; 26: 3563-3574) and maternal metabolic status, for example, diabetes (N Hajj et al., Diabetes. CIO: 10.2337/db12-0289) have been associated with epigenetic aberrations in offspring.

Despite decades of work, no practical means were available to study fetal or tumor methyloma and to monitor dynamic changes throughout pregnancy or during disease processes such as malignancies. Thus, there is a need for non-invasive methods for the analysis of complete fetal and tumor methyloma or fragments thereof.

- 1 042157

Brief description of the invention

Embodiments provide systems, methods, and devices for determining and applying methylation profiles of various tissues and samples. Examples are given. A methylation profile can be inferred for fetal/tumor tissue by comparing plasma methylation (or other sample containing cell-free DNA, e.g. urine, saliva, genital swabs) with the mother/patient's methylation profile. The methylation profile can be determined for fetal/tumor tissue using tissue-specific alleles to identify fetal/tumor DNA if the sample contains a mixture of DNA. The methylation profile can be used to determine copy number variation in the fetal/tumor genome. Fetal methylation markers have been identified using various techniques. The methylation profile can be determined by determining the size distribution of the size of the DNA fragments, while the reference values of the size parameter can be used to determine the levels of methylation.

In addition, the level of methylation can be used to determine the level of cancer. In the context of cancer, measurement of changes in plasma methylome may allow cancer detection (eg, for screening purposes), monitoring (eg, detecting response to cancer therapy; and detecting cancer recurrence), and predictive (eg, measuring body burden of cancer cells, or for staging, assessing the likelihood of death as a result of the disease, progression of the disease or metastatic processes).

A better understanding of the nature and advantages of the embodiments of the present invention can be achieved on the basis of the following detailed description and accompanying drawings.

Brief description of the drawings

In FIG. 1A shows the table. 100 sequencing results for maternal blood, placenta, and maternal plasma according to embodiments of the present invention.

In FIG. 1B shows the methylation density of 1 Mb patches from sequenced samples in accordance with embodiments of the present invention.

In FIG. 2A-C are graphs of beta values versus methylation indices: (A) Maternal blood cells, (B) Chorionic villus sample, (C) Mature placental tissue.

In FIG. 3A and B are histograms of the percentage of methylated CpG sites in plasma and blood cells from adult male and non-pregnant adult female: (A) Autosomes, (B) X chromosome.

In FIG. 4A and B are graphs of the methylation densities of the respective loci in blood cell DNA and plasma DNA: (A) Non-pregnant adult female, (B) Adult male.

In FIG. 5A and B are histograms of the percentage of methylated CpG sites in pregnancy samples: (A). Autosomes, (B) X chromosome.

In FIG. Figure 6 shows a histogram of the level of methylation of different classes of repeats in the human genome for maternal blood, placenta, and maternal plasma.

In FIG. 7A shows a Circos plot 700 for first trimester samples.

In FIG. 7B shows a Circos plot 750 for samples obtained in the third trimester.

In FIG. 8A-D are graphs comparing genomic tissue DNA methylation densities with maternal plasma DNA for CpG sites surrounding informative single nucleotide polymorphisms.

Fig. 9 is a flowchart illustrating a method 900 for determining a first methylation profile from a biological sample of an organism, in accordance with embodiments of the present invention.

Fig. 10 is a flowchart illustrating a method 1000 for determining a first methylation profile from a biological sample of an organism, in accordance with embodiments of the present invention.

In FIG. 11A and B are graphical representations of the performance of a prediction algorithm using maternal plasma data and fetal DNA fractional concentration in accordance with embodiments of the present invention.

Fig. 12A is a table. 1200 for details of 15 selected genomic loci for predicting methylation according to embodiments of the present invention.

In FIG. 12B is a graphical representation of the 1250 inferred categories for 15 selected genomic loci and their corresponding placental methylation levels.

Fig. 13 is a flow diagram of a method 1300 for detecting a fetal chromosomal abnormality from a biological sample obtained from a subject who is pregnant with at least one fetus.

Fig. 14 is a flow diagram of a method 1400 for identifying methylation markers by comparing placental methylation profile with maternal methylation profile, in accordance with embodiments of the present invention.

Fig. 15A is a table 1500 representing the performance of the identification algorithm.

- 2 042157 DMR cations using data for the first trimester, taking into account 33 previously described markers of the first trimester.

Fig. 15B is a table 1550 showing the performance of the DMR identification algorithm using third trimester data compared to a placental sample obtained at delivery.

Fig. 16 is a table. 1600, which shows the numbers of loci hypermethylated or hypomethylated predicted based on direct analysis of maternal plasma bisulfite sequencing data.

Fig. 17A is a graph 1700 showing the size distribution of maternal plasma DNA, non-pregnant control plasma, placental DNA, and peripheral blood DNA.

Fig. 17B is a plot 1750 of the size distribution and methylation profile of maternal plasma, adult female control plasma, placental tissue, and adult female control blood.

Fig. 18A and B are plots of methylation densities and plasma DNA molecular sizes in accordance with embodiments of the present invention.

In FIG. 19A is a plot of 1900 methylation densities and sequencing read sizes for an adult non-pregnant woman.

Fig. 19B is a 1950 graph showing the size distribution and methylation profile of fetal-specific and maternal-specific DNA molecules in maternal plasma.

Fig. 20 is a flow diagram of a method 2000 for assessing the level of DNA methylation in a biological sample of an organism, in accordance with embodiments of the present invention.

Fig. 21A is a table. 2100 for methylation densities for preoperative plasma and tissue samples from a patient with hepatocellular carcinoma (HCC).

Fig. 21B is a table. 2150 for the number of sequence rows and the achieved sequencing depth per sample.

Fig. 22 is a table. 220, which shows autosomal methylation densities ranging from 71.2 to 72.5% in plasma samples from healthy controls.

In FIG. 23A and B show the methylation density of buffy coat, tumor tissue, non-tumor liver tissue, preoperative plasma, and postoperative plasma from a patient with HCC.

Fig. 24A is a graph 2400 showing the methylation densities of preoperative HCC patient plasma.

Fig. 24B is a graph 2450 showing the methylation densities of postoperative plasma from an HCC patient.

In FIG. 25A and B are Z-scores of plasma DNA methylation densities for preoperative (plot 2500) and postoperative (plot 2550) plasma samples from an HCC patient using plasma methylome data from four healthy controls as a reference for chromosome 1.

Fig. 26A is a table. 2600 for preoperative and postoperative plasma Z-score data.

Fig. 26B is a Circos plot 2620 plotting the Z-score of plasma DNA methylation densities for preoperative and postoperative HCC patient plasma samples using four healthy controls as a reference for 1 Mb patches analyzed for all autosomes.

Fig. 26C is a table. 2640 representing the Z-score distribution of 1 Mb regions for the whole genome in both preoperative and postoperative plasma samples from a patient with HCC.

Fig. 26D is a table. 2660, which presents methylation levels in a sample of tumor tissue and preoperative plasma, overlapping with some of the control plasma samples in the context of HH and CHG.

In FIG. 27A-H are Circos plots of methylation density in 8 cancer patients according to embodiments of the present invention.

Fig. 27I is Table 2780 listing the number of sequence reads and the achieved sequencing depth per sample.

Fig. 27J is a table 2790 showing the Z-score distribution of 1 Mb regions for the whole genome in the plasma of patients with various malignancies. AL = lung adenocarcinoma; NHC = nasopharyngeal carcinoma; CRC = colorectal carcinoma; NEC = neuroendocrine carcinoma; HMS = smooth muscle sarcoma.

Fig. 28 is a flowchart of a method 2800 for analyzing a biological sample of an organism to establish a cancer level classification, in accordance with embodiments of the present invention.

Fig. 29A is a graph 2900 showing the distribution of methyl densities

- 3 042157 reference subjects under the assumption that the specified distribution corresponds to the normal distribution.

Fig. 29B is a graph 2950 showing the distribution of methylation densities in cancer patients, assuming that the distribution follows a normal distribution, and the mean methylation level is 2 standard deviations below the cutoff.

Fig. 30 is a graph 3000 showing the distribution of plasma DNA methylation densities in healthy subjects and cancer patients.

In FIG. 31 is a graphical representation 3100 of the distribution of differences in mean DNA methylation densities between plasma of healthy subjects and tumor tissue from an HCC patient.

Fig. 32A is Table 3200 showing the effect of decreasing sequencing depth in cases where the plasma sample contained 5% or 2% tumor DNA.

In FIG. 32B is a graphical representation of 3250 methylation densities of repeat elements and non-repeat regions in the plasma of four healthy controls, buffy coat, normal liver tissue, tumor tissue, preoperative plasma, and postoperative plasma from an HCC patient.

In FIG. 33 is a block diagram of an example computer system 3300 suitable for use with the system and methods in accordance with embodiments of the present invention.

In FIG. 34A shows the size distribution of plasma DNA in a patient with systemic lupus erythematosus (SLE) SLE04.

In FIG. 34B and C are methylation analyzes for plasma DNA from the SLE patient SLE04 (FIG. 34B) and the HCC patient TBR36 (FIG. 34C).

Fig. 35 is a flowchart of a method 3500 for establishing cancer grade classification based on CpG island hypermethylation, in accordance with embodiments of the present invention.

Fig. 36 is a flow diagram of a method 3600 for analyzing a biological sample of an organism using multiple chromosomal regions, in accordance with embodiments of the present invention.

In FIG. 37A shows CNA (copy number aberration) analysis for tumor tissues, bisulfite-untreated (BS) plasma DNA, and bisulfite-treated plasma DNA (center-to-edge) for patient TBR36.

Fig. 37B is a scatter plot plotting the relationship between Z-scores for CNA detection using bisulfite-treated and non-bisulfite-treated plasma for 1 Mb patches in patient TBR36.

In FIG. 38A shows CNA analysis for tumor tissues, bisulfite-free (BS) plasma DNA and bisulfite-treated plasma DNA (center to edges) for patient TBR34.

Fig. 38B is a scatter plot plotting the relationship between Z-scores for CNA detection using bisulfite-treated and non-bisulfite-treated plasma for 1 Mb patches in patient TBR34.

Fig. 39A is a Circos plot showing CNA (inner circle) and methylation analysis (outer circle) for bisulfite-treated plasma from a TBR240 HCC patient.

Fig. 39B is a Circos plot showing CNA (inner circle) and methylation analysis (outer circle) for bisulfite-treated plasma from a TBR164 HCC patient.

In FIG. 40A shows CNA analysis of a TBR36 patient pre-treatment sample and a post-treatment sample.

In FIG. 40B shows methylation analysis of a TBR36 patient in a pre-treatment sample and a post-treatment sample.

In FIG. 41A shows the CNA analysis of the patient TBR34 pre-treatment sample and post-treatment sample.

Fig. 41B shows methylation analysis in a TBR34 patient of a pre-treatment sample and a post-treatment sample.

In FIG. 42 is a diagram of the diagnostic performance of genome-wide hypomethylation analysis with different numbers of sequencing reads.

Fig. 43 is a graph showing ROC curves for cancer detection based on genome-wide hypomethylation analysis at different plot sizes (50 kb, 100 kb, 200 kb and 1 Mb).

In FIG. 44A shows the diagnostic performance of the sum probability (CP) and the percentage of areas with aberrations.

In FIG. 44B shows the diagnostic performance of a plasma assay to assess global hypomethylation, CpG islands of hypermethylation, and CNA.

In FIG. 45 is a table of results for global hypomethylation, CpG islands of hypermethylation, and CNA in patients with hepatocellular carcinoma.

In FIG. 46 is a table of results for global hypomethylation, CpG islands of hypermethylation, and CNA in patients with cancers other than hepatocellular.

- 4 042157 carcinomas.

In FIG. 47 shows a serial analysis of plasma methylation for patient TBR34.

In FIG. 48A is a Circos plot showing changes in CNA (inner circle) and methylation (outer circle) in DNA bisulfite-treated plasma from a TBR36 HCC patient.

Fig. 48B is a plot of methylation Z-scores for regions with addition or loss of chromosomal material and regions without change in copy number for a TBR36 HCC patient.

In FIG. 49A is a Circos plot showing changes in CNA (inner circle) and methylation (outer circle) in bisulfite-treated plasma DNA from a TBR34 HCC patient.

Fig. 49B is a plot of methylation Z-scores for areas with addition or loss of chromosomal material and areas without change in copy number for a TBR34 HCC patient.

In FIG. 50A and B show the results of plasma hypomethylation and CNA analysis for patients with CKB, CKB04, and CKB10.

In FIG. 51A and B are Z _meth analyzes for CNA and non-CNA regions in the plasma of two HCC patients (TBR34 and TBR36).

Fig. 51C and D are Z _meth analyzes for areas with and without CNA in the plasma of two CKB patients (CKB04 and CKB10).

In FIG. 52A shows a hierarchical cluster analysis for plasma samples from HCC patients, non-HCC patients, and healthy controls using group A traits for CNA, global methylation, and CpG island methylation.

In FIG. 52B shows hierarchical clustering using group B traits for CNA, global methylation, and CpG island methylation.

In FIG. 53A shows a hierarchical cluster analysis for plasma samples from HCC patients, patients with non-HCC disease, and healthy controls using group A CpG island methylation signatures.

In FIG. 53B shows a hierarchical cluster analysis of plasma samples from HCC patients, patients with non-HCC disease, and healthy controls using group A global methylation densities.

In FIG. 54A shows a hierarchical cluster analysis of plasma samples from HCC patients, patients with non-HCC disease, and healthy controls using the global group A CNA content.

In FIG. 54B shows a hierarchical cluster analysis of plasma samples from HCC patients, patients with non-HCC disease, and healthy controls using the methylation densities of group B CpG islands.

In FIG. 55A shows a hierarchical cluster analysis for plasma samples from HCC patients, patients with non-HCC disease, and healthy controls using group B global methylation densities.

In FIG. 55B shows a hierarchical cluster analysis of plasma samples from HCC patients, patients with non-HCC disease, and healthy controls using group B global methylation densities.

In FIG. 56 shows the mean methylation density of 1 Mb patches (red dots) in 32 healthy subjects.

Definitions

Methylome is a measure of the amount of DNA methylation at multiple sites or loci in the genome. A methylome can correspond to the entire genome, a substantial part of the genome, or relatively small(s) part(s) of the genome. The fetal methylome corresponds to the fetal methyloma of a pregnant woman. Fetal bromide can be determined using a variety of fetal tissues or sources of fetal DNA, including placental tissues and cell-free fetal DNA in maternal plasma. The tumor methylome corresponds to the methylome of the tumor of an organism (for example, a human). Tumor methyloma can be determined using tumor tissue or cell-free tumor DNA in maternal plasma. Fetal methylome and tumor methylome are examples of methylome of interest. Other examples of methylomes of interest are organ methylomes (e.g., methylomes of brain, bone, lung, heart, muscle, and kidney cells, etc.) from which DNA can be released into body fluids (e.g., plasma, serum, sweat, saliva, urine, genital secretions, seminal fluid, stool fluid, diarrheal fluid, cerebrospinal fluid, gastrointestinal secretions, pancreatic secretions, intestinal secretions, sputum, tears, aspiration fluid from the breast and thyroid gland, etc. .d.). Said organs may be transplanted organs.

Plasma methylome is methylome as determined from animal (eg, human) plasma or serum. Plasma methylome is an example of cell-free DNA methylome because plasma or serum contains cell-free DNA. Plasma methylome is also an example of mixed methylome because it is a mixture of fetal/maternal

- 5 042157 methyloma or tumor/patient methyloma. Placental methylome can be determined from a chorionic villus sample (CVS) or a sample of placental tissue (eg, obtained after childbirth). The cellular methylome corresponds to the methylome determined from the cells (eg, blood cells) of the patient. Methyl in blood cells is called blood cell methylome (or blood methylome).

A site corresponds to a single site, which may be a single base or a group of related bases, such as a CpG site. A locus may correspond to an area that includes multiple sites. A locus can include just one site, making a locus equivalent to a site in this context.

The methylation index of each genomic site (eg, CpG site) refers to the ratio of sequencing reads showing methylation at said site to the total number of reads covering said site. The methylation density of a region is the number of reads showing methylation at sites within the specified region divided by the total number of reads covering sites in the specified region. The sites may have specific characteristics, such as being CpG sites. Accordingly, the CpG methylation density for a region is the number of reads showing CpG methylation divided by the total number of reads covering CpG sites in the region (e.g., a specific CpG site, CpG sites within a CpG island, or a larger region). For example, the methylation density for each 100-kb region in the human genome can be determined from the total number of cytosines not converted after bisulfite treatment (corresponding to methylated cytosine) into CpG sites, as a proportion of all CpG sites covered by sequence reads mapped to 100 kb area. This analysis can also be carried out for sections of other sizes, for example, 50 kb or 1 Mb, etc. The region may be the entire genome, or a chromosome, or a portion of a chromosome (eg, a chromosome arm). The methylation index of a CpG site matches the methylation density of a region if the specified region includes only the specified CpG site. The proportion of methylated cytosines refers to the ratio of the number of cytosine sites, C, that are shown to be methylated (eg not converted after treatment with bisulfite) to the total number of cytosine residues analyzed, i.e. including cytosines outside the context of CpG, in the specified area. The methylation index, methylation density, and proportion of methylated cytosines are examples of methylation levels.

The methylation profile (also called methylation status) includes information on DNA methylation for a region. DNA methylation information may include, but is not limited to, CpG site methylation index, methylation density of CpG sites in a region, distribution of CpG sites over a contiguous region, methylation pattern or level for each individual CpG site within a region that contains more than one CpG site, and methylation at non-CpG sites. The methylation profile of a significant part of the genome can be considered equivalent to a methylome. DNA methylation in mammalian genomes generally refers to the addition of a methyl group to the 5' carbon of cytosine residues (ie, 5-methylcytosines) in CpG dinucleotides. DNA methylation at cytosines can occur in other contexts, for example, in CHG and CHH, where H is adenine, cytosine, or thymine. Cytosine methylation can also be represented by 5hydroxymethylcytosine. Methylation at non-cytosine bases, such as K6-methyladenine, has also been described.

The fabric corresponds to any cells. Different tissue types may correspond to different cell types (eg, liver, lung, or blood) and may also correspond to tissues from different organisms (mother/fetus) or healthy cells/tumor cells. A biological sample refers to any sample that is obtained from a subject (e.g., a human such as a pregnant woman, an individual with or suspected of having cancer, an organ transplant recipient, or a subject suspected of having a disease process in which an organ is involved (for example, the heart in myocardial infarction, or the brain in stroke) and contains one or more nucleic acid molecules of interest.The biological sample may be a body fluid such as blood, plasma, serum, urine, vaginal fluid, washings from the uterus or vagina, pleural fluid, ascitic fluid, cerebrospinal fluid, saliva, sweat, tears, sputum, bronchoalveolar lavage fluid, etc. Stool samples may also be used.

The term cancer level can refer to both the presence of cancer and the stage of cancer, tumor size, presence of metastases, total tumor burden on the body, and/or other measure of cancer severity. The level of cancer can be expressed as a number or otherwise indicated. The level may be 0. The level of cancer also includes premalignant or precancerous conditions (statuses) associated with mutations or a number of mutations. The cancer rate can be used in a variety of ways. For example, screening can indicate whether someone who is not known to have previously had cancer has cancer. Assessment may be done on someone diagnosed with cancer to monitor progression

- 6 042157 cancer over time, to study the effectiveness of therapy or to determine the prognosis. In one embodiment, the prognosis may be represented as the probability of a patient dying from a cancer, or the probability of the cancer progressing after a certain period or time interval, or the probability of the cancer metastasizing. Detection may mean screening, or may mean checking for cancer in someone showing signs (eg, symptoms or other positive tests) suggestive of cancer.

Detailed description of the invention

Epigenetic mechanisms play an important role in the development of the embryo and fetus. However, human embryonic and fetal tissues (including placental tissues) are not readily available (US Pat. No. 6,927,028). Some embodiments address this problem by analyzing a sample containing a cell-free fetal DNA molecule present in the maternal circulation. Fetal bromide can be bred in a variety of ways. For example, maternal plasma methyloma and maternal cellular methyloma (by maternal blood cells) can be compared and differences found that correlate with fetal methyloma. In another example, fetal-specific alleles can be used to determine fetal methylome methylation at specific loci. In addition, fragment size can be used as an indicator of percent methylation, as a correlation between size and percent methylation is shown.

In one embodiment, whole genome bisulfite sequencing is used to analyze the methylation profile (of all or part of the methylome) of maternal plasma DNA with single nucleotide resolution. Using differences in maternal and fetal polymorphisms, fetal bromide can be collected from maternal blood samples. In another embodiment, polymorphism differences were not used, but a distinction between plasma methylome and blood cell methylome could be used.

In another embodiment, by using single nucleotide variations and/or copy number aberrations between the tumor genome and the non-tumor genome, as well as plasma (or other sample) sequestration data, a tumor methylation profile can be obtained for a sample from a patient who is suspected or certain to have has cancer. The difference in the level of methylation in the plasma sample of the test individual when compared with the level of methylation in the plasma of a healthy control or group of healthy controls can identify the test individual as having cancer. In addition, the methylation profile can serve as a signature of the type of cancer, for example, it can show from which organ the cancer originates in a given individual, and whether there are metastases.

Due to the non-invasive nature of this method, the present inventors were able to serially evaluate fetal and maternal plasma methylomes from maternal blood samples collected during the first trimester, third trimester, and postpartum. There were changes associated with gestation. This method can also be applied to samples obtained during the second trimester. Fetal bromide, as measured by maternal plasma during pregnancy, resembled placental bromide. Imprinting genes and differentially methylated regions were identified from maternal plasma data.

Thus, the authors have developed a method for a serial, comprehensive non-invasive non-invasive study of fetal methyloma, offering, respectively, the possibility of identifying biomarkers or direct testing associated with pathologies of pregnancy. Embodiments may also be used for serial, comprehensive, non-invasive testing of tumor methyloma, for screening or detecting whether a subject has cancer, for monitoring malignancies in a cancer patient, and for prognostication. Embodiments may be applied to any type of cancer including, but not limited to, lung cancer, breast cancer, colon and rectal cancer, prostate cancer, nasopharyngeal cancer, gastric cancer, testicular cancer, skin cancer (e.g., melanoma ), cancer affecting the nervous system, bone cancer, ovarian cancer, liver cancer (eg, hepatocellular carcinoma), hematological malignancies, pancreatic cancer, endometrioid carcinoma, kidney cancer, cervical cancer, bladder cancer, etc.

Methods for determining the methylome or methylation profile are discussed first, then the various methylomas (such as fetal methylomas, tumor methylomas, maternal or patient methylomas, and mixed methylomas, eg plasma) are described. Determination of the fetal methylation profile is then described using fetal-specific markers or by comparing the mixed methylation profile with the cellular methylation profile. Fetal methylation markers are determined by comparing methylation profiles. The relationship between size and methylation is discussed. Also proposed options for the use of methylation profiles for the detection of cancers.

I. Determination of methylome

Many approaches have been used to study placental methyloma, but each

- 7 042157 approach has its limitations. For example, sodium bisulfite, a chemical that modifies unmethylated cytosine residues to uracil and leaves methylated cytosine unchanged, converts cytosine methylation differences into genetic sequence differences for further study. The gold standard for studies of cytosine methylation is based on the treatment of tissue DNA with sodium bisulfite followed by direct sequencing of individual clones of bisulfite-converted DNA molecules. After analysis of multiple clones of DNA molecules, the cytosine methylation pattern and quantitative profile per CpG site can be obtained. However, bisulfite sequencing of clones is a labor intensive, low throughput procedure that cannot be easily applied at the whole genome scale.

Methylation-sensitive restriction enzymes, which typically cleave unmethylated DNA, provide an inexpensive way to study DNA methylation. However, the data provided by such studies are limited to loci containing enzyme-recognized motifs, and the results are not quantitative. Immunoprecipitation of DNA bound by antibodies against methylated cytosine can be used to study large segments of the genome, but it tends to be biased in favor of loci with high methylation density due to more intense antibody binding to such regions. Microarray-based methods depend on the original design of the assay probes and the efficiency of the probes to hybridize to the target DNA.

For a comprehensive study of methylome, in some embodiments, massively parallel sequencing (MPS) is used to obtain genome-wide information and quantify the level of methylation per nucleotide and per allele. Recently, bisulfite conversion followed by genome-wide MPS has become possible (R Lister et al 2008 Cell; 133: 523-536).

From a small number of published studies (R Lister et al. 2009 Nature; 462: 315-322; L Laurent et al. 2010 Genome Res; 20: 320-331; Y Li et al. 2010 PLoS Biol; 8: e1000533; and M Kulis et al. 2012 Nat Genet; 44: 1236-1242), where whole genome bisulfite sequencing has been used to investigate human methylomes, two studies focus on embryonic stem cells and fetal fibroblasts (R Lister et al. 2009 Nature; 462: 315-322; L Laurent et al 2010 Genome Res 20: 320331). Both studies examined DNA derived from cell lines.

A. Whole genome bisulfite sequencing.

Some implementation options allow you to overcome the above problems and conduct a comprehensive non-invasive serial study of fetal methyloma. In one embodiment, whole genome bisulfite sequencing was used to analyze cell-free fetal DNA molecules present in the bloodstream of pregnant women. Despite the low content and fragmented nature of DNA molecules in plasma, the authors were able to collect high-resolution fetal methyloma from maternal plasma and conduct serial observations of changes during pregnancy. Given that there is significant interest in non-invasive prenatal testing (NIPT), implementations may provide a powerful new tool for searching for fetal biomarkers or serve directly as a platform for conducting NIPT for fetal or pregnancy-related diseases. The data of whole genome bisulfite sequencing of various samples are presented, based on which fetal methyloma can be established. In one embodiment, this technology can be used to determine the methylation profile during pregnancy complicated by preeclampsia, or intrauterine growth retardation, or preterm birth. In such complicated pregnancies, this technology can be used serially due to its non-invasive nature, allowing monitoring and/or prediction and/or evaluation of response to treatment.

In FIG. 1A shows the table. 100 sequencing results for maternal blood, placenta, and maternal plasma according to embodiments of the present invention. In one embodiment, whole genome sequencing was performed on bisulfite-converted DNA libraries prepared using methylated DNA library adapters (Illumina) (R Lister et al. 2008 Cell; 133: 523-536), blood cells from a first trimester blood sample. , CVS, placental tissue collected at term, maternal plasma samples collected in the first and third trimesters and in the postpartum period. Blood and plasma DNA samples from one adult male and one adult non-pregnant woman were also examined. In total, 9.5 billion pairs of raw sequence reads were obtained in this study. The sequencing depth for each sample is shown in Table 1. 100.

Sequence reads uniquely mapped to the human reference genome provided 50-fold, 34-fold, and 28-fold mean haploid genome coverage, respectively, for maternal plasma samples in the first trimester, third trimester, and postnatal period. Coverage of CpG sites in the genome varied from 81 to 92% for samples obtained during pregnancy. Sequence reads capturing CpG sites provided 33-fold average haploid genome coverage per strand, 23-fold coverage per strand, and 19-fold coverage per strand, respectively, for pattern

- 8 042157 maternal plasma in the first trimester, third trimester and postpartum period. Bisulfite conversion efficiency for all samples was >99.9% (Table 100).

In table. The 100 ambiguity score (marked with index a) refers to the proportion of reads mapped to both Watson-Crick strands (both the 5'^3' strand and the 3'^5' strand) of the reference human genome. The lambda conversion index refers to the proportion of unmethylated cytosines in the internal control, lambda DNA, converted to thymine residues by bisulfite modification. H is a generalized designation for A, C, or T. a refers to reads that can be mapped to a specific genomic locus but cannot be mapped to a Watson (5'^3') or Crick (3'^5') chain. b refers to paired reads with identical start and end coordinates, c: some amount of lambda DNA was added to each sample prior to bisulfite conversion. The lambda conversion index refers to the proportion of cytosine nucleotides that remain cytosine after bisulfite conversion and is used as an indicator of the rate of successful bisulfite conversion, d refers to the number of cytosine nucleotides present in the human reference genome that retain the cytosine sequence after bisulfite conversion.

With bisulfite modification, unmethylated cytosines are converted to uracils and subsequently to thymines after PCR amplification, while methylated cytosines remain intact (M Frommer et al. 1992 Proc Natl Acad Sci USA; 89:1827-31). After sequencing and alignment, the methylation status of an individual CpG site can, respectively, be inferred by counting methylated M (methylated) sequence reads and unmethylated U (unmethylated) sequence reads by cytosine residues in the CpG context. Using bisulfite sequencing data, complete methylomas of maternal blood, placenta, and maternal plasma were constructed. The mean CpG methylation density (also called MD methylation density) of specific loci in maternal plasma can be calculated using the equation:

M

MD =--M + U where M is the counted number of methylated reads and U is the counted number of unmethylated reads at CpG sites within the genetic locus. If the locus contains more than one CpG site, M and U correspond to the counted number in all these sites.

B. Various techniques.

As described above, methylation profiling can be performed using massively parallel sequencing (MPS) of bisulfite-converted plasma DNA. MPS of bisulfite-converted DNA plasma can be performed randomly or by the shotgun method. The depth of sequencing may vary according to the size of the region of interest.

In another embodiment, the region(s) of interest in bisulfite-converted plasma DNA may be first captured(s) using a liquid phase or solid phase hybridization method followed by MPS. Massively parallel sequencing can be performed using synthesis sequencing platforms such as Illumina, ligation sequencing platforms such as the SOLiD platform from Life Technologies, semiconductor-based sequencing systems such as the Ion Torrent or Ion Proton platforms from Life Technologies, or a single molecule sequencing system such as Helicos or Pacific Biosciences, or a nanopore based sequencing system. Sequencing based on nanopores, including nanopores designed using, for example, lipid bilayers, protein nanopores, and solid-state nanopores (eg, based on graphene). Because the selected single-molecule sequencing platforms allow the methylation of DNA molecules (including N6methyladenine, 5-methylcytosine, and 5-hydroxymethylcytosine) to be determined directly, without bisulfite conversion (BA Flusberg et al. 2010 Nat Methods; 7: 461-465; J Shim et al. 2013 Sci Rep 3:1389 CIO: 10.1038/srep01389), the use of such platforms allows analysis of the methylation status of a sample of bisulfite-unconverted DNA (eg, plasma DNA).

Along with sequencing, other techniques can be used. In one embodiment, methylation profiling can be performed using methylation-specific PCR or methylation-sensitive restriction enzyme digestion followed by PCR or ligase chain reaction followed by PCR. In other embodiments, PCR is a form of single molecule or digital PCR (V Vogelstein et al. 1999 Proc Natl Acad Sci USA; 96: 9236-9241). In additional embodiments, the PCR may be real-time PCR. In other embodiments, the PCR may be multiplex PCR.

II. Methylome analysis

Some implementation options allow you to determine the profile of plasma DNA methylation using whole genome bisulfite sequencing. The fetal methylation profile can

- 9 042157 be determined by DNA sequencing from maternal plasma samples as described below. Accordingly, fetal DNA molecules (and fetal methylome) were assessed in a non-invasive manner during pregnancy and changes were monitored serially as the pregnancy progressed. Due to the completeness of sequencing data, the authors were able to study maternal plasma methylomes on a genome-wide scale with a resolution of up to one nucleotide.

Since the genomic coordinates of the sequencing reads were known, these data allowed us to explore the overall methylation levels of methylome or any region of interest in the genome, and to make comparisons between different genetic elements. In addition, each CpG site or locus was covered by several sequence reads. This document describes some of the metrics used to measure methylome.

A. Methylation of plasma DNA molecules.

DNA molecules are present in human plasma at low concentrations, in fragments that generally resemble mononucleosomal units in length (YMD Lo et al. 2010 Sci Transl Med; 2: 61ra91; and YW Zheng et al. 2012 Clin Chem; 58: 549-558). Despite these limitations, the whole genome bisulfite sequencing system made it possible to analyze the methylation of plasma DNA molecules. In other embodiments, since the selected single-molecule sequencing platforms allow the methylation status of DNA molecules to be determined directly without bisulfite conversion (BA Flusberg et al. 2010 Nat Methods; 7: 461-465; J Shim et al. 2013 Sci Rep; 3: 1389. CIO: 10.1038/srep01389), the use of such platforms allows the use of bisulfite-unconverted plasma DNA to determine plasma DNA methylation levels or to determine plasma methylome. Such platforms allow the detection of N6-methyladenine, 5-methylcytosine, and 5hydroxymethylcytosine, which may provide improved results (eg, improved sensitivity or specificity) for different biological functions of different methylation forms. Such improved results may be useful when implementing embodiments for the detection or monitoring of specific disorders, such as pre-eclampsia or a certain type of cancer.

Bisulfite sequencing also makes it possible to distinguish between different forms of methylation. In one embodiment, additional steps may be included to distinguish between 5-methylcytosine and 5-hydroxymethylcytosine. One such approach is oxidative bisulfite sequencing (oxBS-seq), which can localize 5-methylcytosine and 5hydroxymethylcytosine with one base resolution (MJ Booth et al. 2012 Science; 336: 934-937; MJ Booth et al. 2013 Nature Protocols 8: 1841-1851). In bisulfite sequencing, both 5-methylcytosine and 5-hydroxymethylcytosine are read as cytosines and therefore cannot be distinguished. On the other hand, in oxBS-seq, specific oxidation of 5-hydroxymethylcytosine to 5-formylcytosine by treatment with potassium perruthenate (KRuO4) followed by conversion of the newly formed 5-formylcytosine to uracil using bisulfite conversion makes it possible to distinguish 5hydroxymethylcytosine from 5-methylcytosine. Thus, registration of 5-methylcytosine can be performed in one cycle of oxBS-seq, and the levels of 5-hydroxymethylcytosine are derived by comparison with the results of bisulfite sequencing. In another embodiment, 5-methylcytosine can be distinguished from 5-hydroxymethylcytosine using bisulfite sequencing using Tet (TAB-seq) (M Yu et al. 2012 Nat Protoc; 7: 2159-2170). TAB-seq makes it possible to identify 5hydroxymethylcytosine with a resolution of one nucleotide, as well as to determine its content in each modification site. This method involves β-glucosyltransferase-mediated protection of 5-hydroxymethylcytosine (glucosylation) and recombinant mouse Tet1 (mTet1)-mediated oxidation of 5-methylcytosine to 5-carboxycytosine. After subsequent treatment with bisulfite and PCR amplification, both cytosine and 5-carboxycytosine (derived from 5-methylcytosine) are converted to thymine (T), with 5-hydroxymethylcytosine reading as C.

In FIG. 1B shows the methylation density in 1 Mb sequenced sample regions according to embodiments of the present invention. Plot 150 is a Circos plot showing the methylation density in maternal plasma DNA and genomic DNA in 1 Mb regions throughout the genome. From the center to the edges: chromosome ideograms can be oriented from the short arm to the long arm (pter-qter) clockwise (centromeres are marked in red), maternal blood (red), placenta (yellow), maternal plasma (green), common reads in maternal plasma (blue), and fetal-specific reads in maternal plasma (purple). The total CpG methylation levels (i.e., density levels) of maternal blood cells, placenta, and maternal plasma are shown in Table 1. 100. Maternal blood cell methylation is generally higher than placental methylation across the genome.

B. Comparison of bisulfite sequencing with other techniques.

The authors examined placental methylome using massively parallel bisulfite sequencing. In addition, the authors examined placental methylome using an oligonucleotide array platform covering approximately 480,000 CpG sites in the human genome (Illumina) (M Kulis et al. 2012 Nat Genet; 44: 1236-1242; and C Clark et al. 2012 PLoS One; 7: e50233).

- 10 042157

In one embodiment using BeadChip-based genotyping and methylation analysis, genotyping was performed using an Illumina HumanOmni2.5-8 genotyping array according to the manufacturer's protocol. Genotypes were predicted using the GenCall Genome Studio Software (Illumina) algorithm. Forecasting indicators exceeded 99%. For microarray-based methylation assays, genomic DNA (500-800 ng) was treated with sodium bisulfite using the Zymo EZ DNA Methylation Kit (Zymo Research, Orange, CA, USA) following the manufacturer's recommendations for the Illumina Infinium Methylation Assay.

Methylation analysis was performed on 4 μl of bisulfite-converted genomic DNA at a concentration of 50 ng/μl according to the Infinium HD Methylation Assay methylation assay protocol. The hybridized chip (BeadChip) was scanned using the Illumina iScan tool. DNA methylation data were examined using the GenomeStudio (v2011.1) Methylation Module (v1.9.0) software, with normalization to internal controls and background subtraction. The methylation index for an individual CpG site is presented as a beta value ('), which was calculated using the ratio of the fluorescence intensities of methylated and unmethylated alleles:

_ Density of methylated allele

Unmethylated allele density + Methylated allele density + 100

For CpG sites presented on the matrix and sequenced at least 10-fold in depth, we compared the beta value obtained using the matrix with the methylation index determined by sequencing the same site. Beta values express the density of methylated probes as the ratio of the combined density of methylated and unmethylated probes covering the same CpG site. The methylation index for each CpG site refers to the ratio of methylated reads to the total number of reads covering that CpG.

In FIG. 2A-C are plots of beta values determined using the BeadChip Illumina Infinium HumanMethylation 450K matrix against methylation indices determined using whole genome bisulfite sequencing corresponding to CpG sites examined using both platforms: (A) Maternal blood cells, ( B) Sample of chorionic villi, (C) Tissue of mature placenta. The data from both platforms were in good agreement, with Pearson's correlation coefficients being 0.972, 0.939, and 0.954; R ² values were 0.945, 0.882, and 0.910 for maternal blood cells, CVS, and mature placental tissue, respectively.

The authors also compared their own sequencing data with those described by Chu et al., who examined the methylation profiles of 12 pairs of CVS DNA samples and maternal blood cell DNA using an oligonucleotide template capturing approximately 27,000 CpG sites (T Chu et al. 2011 PLoS One; 6 : e14723). Correlation data between CVS DNA sequencing and maternal blood cells and each of the 12 pairs of samples in the previous study gave mean Pearson's ratio (0.967) and R ² (0.935) for maternal blood, and mean Pearson's ratio (0.943) and R ² (0.888 ) for CVS. The data obtained by the authors of the present invention regarding the CpG sites presented on both matrices correlated well with published data. Methylation levels for non-CpG sites were <1% for maternal blood cells, CVS, and placental tissues (Table 100). These results are consistent with the current notion that significant levels of methylation at non-CpG sites are largely restricted to pluripotent cells (R Lister et al. 2009 Nature; 462: 315-322; L Laurent et al. 2010 Genome Res; 20: 320-331).

C. Comparison of plasma and blood methylomes in non-pregnant subjects.

In FIG. 3A and B are histograms of the percentage of methylated CpG sites in plasma and blood cells from adult male and non-pregnant adult female: (A) Autosomes, (B) X chromosome. Graphs show the similarity of plasma and blood methylomes of a man and a non-pregnant woman. The overall ratios of methylated CpG sites in male and non-pregnant female plasma samples were almost the same as in the corresponding blood cell DNA (Table 100 and FIGS. 2A and B).

The present inventors then investigated the correlation of methylation profiles of plasma samples and blood cells in a locus-specific manner. The authors determined the methylation density of each 100 kb region in the human genome by defining the total number of unconverted cytosines in CpG sites as the proportion of all CpG sites captured by sequence reads mapped to the 100 kb region. The methylation densities were in good agreement in the plasma samples and the corresponding DNA samples of both male and female blood cells.

In FIG. 4A and B are graphs of the methylation densities of the respective loci in blood cell DNA and plasma DNA: (A) Non-pregnant adult female, (B) Adult male. The Pearson correlation coefficient and the R ² value for samples from a non-pregnant woman were 0.963 and 0.927, respectively, and for samples from a man, respectively, 0.953 and 0.908. These data are consistent with previous results based on the evaluation of the genotypes of plasma DNA molecules of recipients of transplanted allogeneic hematopoietic stem cells, which showed that hematopoietic cells are the predominant source of DNA in human plasma (YW Zheng et al. 2012 Clin Chem; 58 : 549-558).

D. Methylation levels of different methylomes.

Next, the present inventors examined the DNA methylation levels of maternal plasma, maternal blood cells, and placental tissue to determine methylation levels. These levels were determined for areas of repetitions, areas without repetitions, and also determined the overall levels.

In FIG. 5A and B are histograms of the percentage of methylated CpG sites in pregnancy samples: (A). Autosomes, (B) X chromosome. The total content of methylated CpG was 67.0% and 68.2% for maternal plasma samples in the first and third trimester, respectively. In contrast to the results obtained from non-pregnant individuals, the indicated content was lower than in the sample of maternal blood cells in the first trimester, but higher than in the samples of CVS and mature placental tissue (Table 100). Notably, the percentage of methylated CpG in a maternal plasma sample from the postpartum period was 73.1%, which is similar to that for blood cells (Table 100). These trends were observed in CpG, occurring in all autosomes, as well as the X chromosome, and covered both areas without repeats and several classes of repeating elements of the human genome.

Both repetitive and non-repetitive elements in the placenta were found to be hypomethylated relative to maternal blood cells. These results were consistent with published data showing that the placenta is hypomethylated relative to other tissues, including peripheral blood cells.

Between 71 and 72% of the sequenced CpG sites were methylated in the DNA of blood cells of a pregnant woman, a non-pregnant woman, and an adult male (Table 100 in Figure 1). These data are comparable with the report by Y Li et al. (Y Li et al. 2010 PLoS Biol; 8: e1000533) about 68.4% CpG sites of blood mononuclear cells. Consistent with previous reports regarding the hypomethylated nature of placental tissues, 55% and 59% of CpG sites were methylated in CVS and mature placental tissue, respectively (Table 100).

In FIG. Figure 6 shows a histogram of the level of methylation of different classes of repeats in the human genome, for maternal blood, placenta, and maternal plasma. These repeat classes correspond to specific UCSC genome browsers. The data for the samples of the first trimester are given. In contrast to previous data suggesting that the hypomethylated nature of placental tissues was observed mainly in certain repeat classes in the genome (In Novakovic et al. 2012 Placenta; 33: 959-970), in the present description, the inventors demonstrate that the placenta at its in fact, it is hypomethylated by most classes of genomic elements relative to blood cells.

E. Similarity of methylomes.

Embodiments allow the determination of methylomes in placental tissues, blood cells and plasma using the same platform. Accordingly, direct comparisons of the methylomes of these types of biological samples could be made. The high level of methylome similarity between male and non-pregnant female blood and plasma cells, and maternal blood cells and postpartum maternal plasma sample further support that hematopoietic cells are the main source of DNA in human plasma (YW Zheng et al. 2012 Clin Chem; 58: 549-558).

The similarity is evident in terms of both the total proportion of methylated CpGs in the genome and the high degree of correlation between the methylation densities of the corresponding loci in blood cell DNA and plasma DNA. However, the total proportion of methylated CpG in maternal plasma samples in the first trimester and in the third trimester was reduced compared to data for maternal blood cells or maternal plasma sample in the postpartum period. Decreased levels of methylation during pregnancy were due to the hypomethylated nature of the fetal DNA molecules present in maternal plasma.

The reversal of the methylation profile in the maternal plasma sample during the postpartum period, becoming more similar to the maternal blood cell profile, suggests that fetal DNA molecules have been eliminated from the maternal circulation. Calculation of fetal DNA concentrations based on fetal SNP markers indeed showed that the indicated concentration changed from 33.9% in the antenatal sample to only 4.5% in the postnatal sample.

F. Other uses.

In embodiments, DNA methylomes have been successfully assembled by MPS analysis of plasma DNA. The ability to detect placental or fetal methyloma from maternal plasma provides a non-invasive way to identify, detect, and monitor aberrant methylation patterns associated with pregnancy-related conditions such as preeclampsia, intrauterine growth retardation, preterm birth, and others. For example, detection of a disease-specific aberrant methylation signature allows screening, diagnosis and monitoring of such pregnancy-related conditions. Measurement of maternal plasma methylation levels allows screening, diagnosis, and monitoring of these pregnancy-related conditions. In addition to direct use for research related to pregnancy

- 12 042157 states, this method can be applied in other areas of medicine for which analysis of plasma DNA is of interest. For example, cancer methylomes can be determined from the plasma DNA of cancer patients. Plasma cancer methylome analysis, as described herein, is a potential synergistic technology for plasma cancer genome analysis (KCA Chan et al. 2013 Clin Chem; 59: 211-224 and RJ Leary et al. 2012 Sci Transl Med; 4 :162ra154).

For example, determination of the methylation level of a plasma sample can be used for cancer screening. If the methylation level of a plasma sample shows aberrant levels compared to healthy controls, the presence of cancer can be assumed. In such a case, additional confirmation and assessment of the type of cancer or tissue of origin of the cancer can be obtained by determining the plasma methylation profile at different genomic loci or by plasma genomic analysis to detect tumor-associated copy number aberrations, chromosomal translocations and single nucleotide variations. Thus, in one embodiment of the present invention, cancer methylome profiling and plasma genome profiling can be performed simultaneously. Alternatively, radiology and imaging (eg, computed tomography, magnetic resonance imaging, positron emission tomography) or endoscopy (eg, upper gastrointestinal endoscopy or colonoscopy) may be used to further investigate individuals who, based on methylation analysis, have plasma is suspected to be cancerous.

In the screening or detection of cancer, the determination of the methylation level of a plasma sample (or other biological material) can be used in combination with other methods of screening or detection of cancer, such as the measurement of prostate specific antigen (for example, for prostate cancer), carcinoembryonic antigen (for example, for colorectal carcinoma, gastric carcinoma, pancreatic carcinoma, lung carcinoma, breast carcinoma, medullary thyroid carcinoma), alpha-fetoprotein (eg, for liver cancer or germ cell tumors), CA125 (eg, for ovarian and breast cancer glands) and CA19-9 (for example, for pancreatic carcinoma).

In addition, other tissues can be sequenced to obtain cellular methylome. For example, liver tissue can be analyzed to determine a liver-specific methylation pattern that can be used to identify liver pathologies. Other tissues that can also be analyzed include brain cells, bones, lungs, heart, muscles and kidneys, etc. The methylation profiles of various tissues may change from time to time, for example, as a result of development, aging, disease processes (eg, inflammation or cirrhosis, or autoimmune processes (such as in systemic lupus erythematosus)) or treatment (eg, treatment with demethylating agents such as 5-azacytidine and 5-azadeoxycytidine). The dynamic nature of DNA methylation makes such an analysis potentially very useful for monitoring physiological and pathological processes. For example, if a change in an individual's plasma methylome from a baseline value obtained when said individual was healthy is detected, disease processes in organs that contribute to plasma DNA can be detected.

The methylomas of transplanted organs can also be determined from the plasma DNA of recipients of transplanted organs. Plasma graft methylome assay according to the disclosure of the present invention is potentially a synergistic technology for plasma graft genomic analysis (YW Zheng et al, 2012; YMD Lo et al. 1998 Lancet; 351: 1329-1330; and TM Snyder et al. 2011 Proc Natl Acad Sci USA; 108: 6229-6234). Since plasma DNA is generally regarded as a marker of cell death, an increase in plasma levels of DNA released from a transplanted organ can be used as a marker of an increase in cell death in that organ, such as a rejection crisis or other pathological processes involving that organ (e.g. , infection or abscess). In the case of successful anti-rejection therapy, the level of DNA released by the transplanted organ in plasma is expected to decrease.

III. Determination of fetal or tumor methyloma using SNPs

As described above, plasma methylome corresponds to blood methylome in a non-pregnant healthy individual. However, in a pregnant woman, these methylomas differ. Fetal DNA molecules circulate in maternal plasma, along with the background majority of maternal DNA molecules (YMD Lo et al. 1998 Am J Hum Genet; 62: 768-775). Thus, in a pregnant woman, plasma methylome is largely a mixture of placental methylome and blood methylome. Accordingly, placental methylome can be obtained using plasma.

In one embodiment, single nucleotide polymorphism (SNP) differences between mother and fetus are used to identify fetal DNA molecules in maternal plasma. The goal was to identify the SNP loci for which the mother is homozygous and the fetus is heterozygous; the fetal-specific allele can be used to determine which DNA fragments belong

- 13 042157 harvest fruit. Genomic DNA from maternal blood cells was examined using the Illumina HumanOmni2.5-8 SNP genotyping matrix. On the other hand, for SNP loci for which the mother is heterozygous and the fetus is homozygous, the mother-specific SNP allele can be used to determine which plasma DNA fragments belong to the mother. The level of methylation of such DNA fragments will reflect the level of methylation of the corresponding genomic regions in the mother.

A. Correlation of methylation of fetal-specific reads and placental methyloma.

Loci containing two different alleles, and the amount of one allele (B) is significantly less than the amount of the other allele (A), were identified by the results of sequencing of a biological sample. Reads covering B alleles were considered fetal-specific (fetal-specific reads). The mother was defined as homozygous for A, and the fetus as heterozygous for A/B, and, accordingly, the reads covering the A allele were common for mother and fetus (common reads).

In one analyzed case of pregnancy, which was used to illustrate several concepts of the present invention, the pregnant mother was found to be homozygous for 1945516 loci on autosomes. Maternal plasma DNA sequencing reads covering these SNPs were studied. Reads carrying a non-maternal allele were found at 107,750 loci, and these loci were considered informative. In each informative SNP, the non-maternal allele was called the fetal-specific allele and the other was called the common allele.

The fractional concentration of fetal/tumor DNA (also called percent fetal DNA) in maternal plasma can be determined. In one embodiment, the fractional concentration of fetal DNA in maternal plasma D is determined from the equation:

where p is the number of sequencing reads with a fetal-specific allele, and q is the number of sequencing reads with a maternal-fetal common allele (YMD Lo et al. 2010 Sci Transl Med; 2: 61ra91). The proportion of fetal DNA in first trimester, third trimester and postpartum maternal plasma samples was found to be 14.4%, 33.9% and 4.5%, respectively. The proportion of fetal DNA was also calculated using the number of reads aligned to the Y chromosome. Based on Y chromosome data, these results were 14.2%, 34.9%, and 3.7%, respectively, in first trimester, third trimester, and postnatal maternal plasma samples.

Embodiments using separate analysis of fetal-specific or common read sequences demonstrate that circulating fetal DNA molecules are significantly more hypomethylated than background DNA molecules. Comparisons of the methylation densities of the respective loci in fetal-specific maternal plasma reads and placental tissue data for both the first and third trimesters showed high levels of correlation. These data provide genomic confirmation that the placenta is the main source of fetal-derived DNA molecules in maternal plasma and represent an important advance over the available loci-based data.

The authors determined the methylation density of each 1 Mb region in the genome using either fetal-specific or general reads capturing CpG sites adjacent to informative SNPs. Fetal and non-fetal specific methylomas assembled from maternal plasma sequence reads can be represented, for example, in a Circos plot (M Krzywinski et al. 2009 Genome Res; 19: 1639-1645). Methylation densities per 1 Mb area were also determined for maternal blood and placental tissue cell samples.

In FIG. 7A shows a Circos plot 700 for first trimester samples. In FIG. 7B shows a Circos plot 750 for samples obtained in the third trimester. Plots 700 and 750 show methylation density per 1 Mb region. Chromosome ideograms (outer circle) are oriented from the short arm to the long arm in a clockwise direction (centromeres are highlighted in red). The second circle from the outside represents the number of CpG sites in the respective 1 Mb areas. For red columns, a scale of up to 20,000 sites per 1 Mb area is given. The methylation densities of the respective 1 Mb regions are shown in other circles using the color scheme shown in the center.

For the first trimester specimens (Fig. 7A), from center to margins, the circles represent: chorionic villus specimen, fetal-specific reeds in maternal plasma, maternal-specific reeds in maternal plasma, combination of fetal and infertile reeds in maternal plasma, and cells maternal blood. For third trimester specimens (Fig. 7B), circles represent: mature placental tissue, fetal-specific reeds in maternal plasma, mother-specific reeds in maternal plasma, a combination of fetal and non-fetal reeds in maternal plasma, in maternal plasma in period and maternal blood cells (from a blood sample of the first trimester). It can be seen that in both the first and third trimester plasma samples, fetal methylomas are more hypomethylated than in non-fetal methylomas.

- 14 042157

The overall methylation profile of fetal methylomas more closely resembled that of CVS or placental tissue samples. On the other hand, the DNA methylation profile for common plasma reads, predominantly representing maternal DNA, more closely resembled that of maternal blood cells. The present inventors then systematically compared the methylation densities of loci one by one, maternal plasma DNA reads, and maternal or fetal tissues. The authors determined the methylation densities of CpG sites that were present on the same sequence read as informative SNPs and that were covered by at least 5 reads of maternal plasma DNA sequences.

In FIG. 8A-D are graphs comparing methylation densities of genomic tissue DNA and maternal plasma DNA for CpG sites surrounding informative single nucleotide polymorphisms. In FIG. 8A shows methylation densities for fetal-specific reads in a first trimester maternal plasma sample versus methylation densities for reads in a CVS sample. As can be seen, the fetal-specific values are in good agreement with the CVS values.

In FIG. 8B shows methylation densities for fetal-specific reads in a third trimester maternal plasma sample versus methylation densities for reads in mature placental tissue. The density sets also agree well, indicating that the fetal methylation profile can be obtained by analyzing reads with fetal-specific alleles.

In FIG. 8C shows methylation densities for total reads in a first trimester maternal plasma sample versus methylation densities for reads in maternal blood cells. Given that most common reads are maternally derived, these two sets of values agree well. In FIG. 8D shows methylation densities for total reads in a third trimester maternal plasma sample versus methylation densities for reads in maternal blood cells.

For maternal plasma fetal-specific reads, the Spearman correlation coefficient between first trimester maternal plasma and CVS was 0.705 (P<2.2*e-16); between third trimester maternal plasma and mature placental tissue, the ratio was 0.796 (P<2.2*e-16) (FIGS. 8A and B). A similar comparison was made for maternal plasma total reads with maternal blood cell data. The Pearson correlation coefficient was 0.653 (P<2.2*e-16) for the first trimester plasma sample, and 0.638 (P<2.2*e-16) for the third trimester plasma sample (Figures 8C and D).

B. Fetal bromide.

In one embodiment, for maternal plasma assembly of fetal methyloma, we selected sequence reads that captured at least one informative fetal SNP site and contained at least one CpG site within the same read. Reads containing fetal-specific alleles were included in the fetal methylome assembly. Reads where a common allele was found, i.e. a non-fetal-specific allele was included in the assembly of a non-fetal-specific methylome, which predominantly consists of maternally derived DNA molecules.

Fetal-specific reads covered 218,010 CpG sites on autosomes in first trimester maternal plasma samples. The corresponding values for third trimester and postpartum maternal plasma samples were 263,611 and 74,020, respectively. On average, total reads covered these CpG sites with an average fold of 33.3, 21.7, and 26.3, respectively. Fetal-specific reads covered these CpG sites at a fold of 3.0, 4.4, and 1.8, respectively, for first trimester, third trimester, and postnatal maternal plasma samples.

Fetal DNA is a minor population in maternal plasma and, accordingly, the coverage of these CpG sites with fetal-specific reads was proportional to the percentage of fetal DNA in the sample. For the maternal plasma sample in the first trimester, the total percentage of methylated CpG in fetal reads was 47.0%, while for total reads it was 68.1%. For the maternal plasma sample in the third trimester, the percentage of methylated CpG in fetal reads was 53.3%, while for total reads it was 68.8%. These data indicated that fetal-specific maternal plasma reeds were more hypomethylated than maternal plasma total reeds.

C. Method.

The techniques described above can also be used to determine the tumor methylation profile. Methods for determining fetal and tumor methylation profiles are described below.

Fig. 9 is a flowchart illustrating a method 900 for determining a first methylation profile from a biological sample of an organism, in accordance with embodiments of the present invention. Method 900 allows the construction of an epigenetic map of the fetus from the maternal plasma methylation profile. The biological sample includes cell-free DNA containing a mixture of cell-free DNA originating from a first tissue and from a second tissue. For example, said first tissue may be from a fetus, from a tumor, or from a transplanted organ.

At step 910, a plurality of DNA molecules from the specified biological sample are analyzed. Analysis of a DNA molecule may include determining the localization of a specified DNA molecule in the genome of a specified

- 15 042157 th organism, determining the genotype of the DNA molecule, and determining whether the specified DNA molecule is methylated at one or more sites.

In one embodiment, DNA molecules are analyzed using reads of DNA sequences, wherein the sequencing is sensitive to methylation. Accordingly, said sequence reads include the methylation status of DNA molecules from said biological sample. The methylation status may include information on whether a particular cytosine residue is 5-methylcytosine or 5-hydroxymethylcytosine. These sequence reads can be obtained using a variety of sequencing techniques, PCR techniques, templates, and other suitable techniques for identifying fragment sequences. The methylation status of sites in a reed sequence can be determined as described herein.

At step 920, a plurality of first loci are identified, characterized in that the first genome of the first tissue is heterozygous for the corresponding first allele, and the corresponding second allele and the second genome of the second tissue are homozygous for the corresponding first allele. For example, fetal-specific reads can be identified at multiple first loci, or tumor-specific reads can be identified at multiple first loci. Tissue-specific reads can be identified by reads in sequencing where the percentage of reads of the second allele sequences falls within a certain range, e.g. approximately 3-25%, thus indicating a minor population of a DNA fragment from a heterozygous genome at the indicated locus and a major population from a homozygous genome at the indicated locus. the specified locus.

At step 930, DNA molecules located at one or more sites of each of the first loci are analyzed. The number of DNA molecules that are methylated at the site and correspond to the corresponding second allele of the specified locus is determined. More than one site per locus may be present. For example, the SNP may indicate that the fragment is specific to the fetus, and the specified fragment may contain several sites, the methylation status of which is determined. The number of methylated reads at each site and the total number of methylated reads for the locus can be determined.

A locus can be defined in terms of a specific number of sites, a specific set of sites, or a specific area size around a variation that contains a tissue-specific allele. A locus can contain only one site. The sites may have specific properties, such as being CpG sites. The determination of the number of unmethylated reads is equivalent to and included in the determination of methylation status.

At step 940, for each of the first loci, the methylation density is calculated based on the numbers of DNA molecules methylated at one or more sites of the specified locus and corresponding to the corresponding second allele of the specified locus. For example, methylation density can be determined for CpG sites corresponding to a locus.

At step 950, a first methylation profile of the first tissue is determined from the methylation densities for the first loci. Said first methylation profile may correspond to specific sites, such as CpG sites. The indicated methylation profile may refer to all loci containing a fetal-specific allele, or only to some of the indicated loci.

IV. Using Differences in Plasma and Blood Methylomes

It has been shown above that fetal-specific plasma reads correlate with placental methyloma. Since the maternal component of maternal plasma methylome is primarily determined by blood cells, the differences between plasma methylome and blood methylome can be used to define placental methylome for all loci, not just the location of fetal-specific alleles. The differences between plasma methylome and blood methylome can also be used to determine tumor methylome.

A. Method.

Fig. 10 is a flowchart illustrating a method 1000 for determining a first methylation profile from a biological sample of an organism, in accordance with embodiments of the present invention. Said biological sample (eg, plasma) comprises cell-free DNA containing a mixture of cell-free DNA originating from a first tissue and from a second tissue. Said first methylation profile corresponds to the methylation profile of the first tissue (eg, fetal tissue or tumor tissue). Method 1200 may provide for the determination of differentially methylated regions from maternal plasma.

In step 1010, a biological sample is obtained. The specified biological sample can be obtained simply with the help of the apparatus (for example, sequencer). Said biological sample may be in a form derived from a body or in a processed form, for example, said sample may be plasma extracted from a blood sample.

At step 1020, a second methylation profile is obtained corresponding to the DNA of the second tissue. Said second methylation profile can be read from memory as it may have been previously determined. Said second methylation profile can be determined from a second tissue, for example, another sample that contains exclusively or predominantly cells from the second tissue. Said second methylation profile may correspond to the cellular methylation profile

- 16 042157 and be derived from cellular DNA. According to another example, said second profile can be determined from a plasma sample obtained before pregnancy or before the development of cancer, since the plasma methylome of a non-pregnant, non-cancer individual is substantially similar to blood cell methylome.

Said second methylation profile may provide methylation density at each of a plurality of loci in the genome of said organism. The density of methylation at a particular locus corresponds to the proportion of methylated DNA in the second tissue. In one embodiment, the methylation density is the methylation density of the CpG, with the CpG sites associated with the locus being used to determine the methylation density. If there is one site per locus, the methylation density can be equal to the methylation index. The density of methylation also corresponds to the density of the absence of methylation, since these two values complement each other.

In one embodiment, said second methylation profile is obtained by performing methylation-sensitive sequencing of cellular DNA from an organism sample. One example of methylation-sensitive sequencing involves treating DNA with sodium bisulfite and then performing DNA sequencing. According to another example, methylation-sensitive sequencing can be performed without the use of sodium bisulfite, using a monomolecular sequencing platform that allows the methylation status of DNA molecules (including N6-methyladenine, 5-methylcytosine and 5-hydroxymethylcytosine) to be determined directly, without bisulfite conversion (AB Flusberg et al 2010 Nat Methods 7: 461-465 J Shim et al 2013 Sci Rep 3:1389 CIO 10.1038/srep01389); or by immunoprecipitation of methylated cytosine (e.g., using an anti-methylcytosine antibody or using a methylated DNA binding protein or peptide (LG Acevedo et al. 2011 Epigenomics; 3: 93-101) followed by sequencing; or by using methylation-sensitive restriction enzymes with in another embodiment, non-sequencing techniques such as arrays, digital PCR, and mass spectrometry are used.

In another embodiment, the second tissue methylation density may be obtained previously from controls from the subject or from other subjects. The methylation density from another subject can work as a reference methylation profile characterized by reference methylation densities. Reference methylation densities can be determined from multiple samples, and the average of the level (or other statistic) of different methylation densities at a locus can be used as a reference methylation density at that locus.

In step 1030, a cell-free DNA methylation profile is determined from the cell-free DNA from the mixture. This methylation profile of cell-free DNA determines the density of methylation at each of the many loci. Said cell-free DNA methylation profile can be determined by obtaining sequence reads from cell-free DNA sequencing, where methylation information is obtained from the sequence reads. This methylation profile of cell-free DNA can be determined in the same manner as cellular methylome.

At 1040, the percentage of cell-free DNA from the first tissue in the biological sample is determined. In one embodiment, said first tissue is fetal tissue and the corresponding DNA is fetal DNA. In another embodiment, said first tissue is tumor tissue and the corresponding DNA is tumor DNA. This percentage can be determined in various ways, for example, using a fetal-specific allele or a tumor-specific allele. The copy number can also be used to determine the indicated percentage, for example, as described in incorporated by reference US patent application 13/801748, entitled Mutational analysis of plasma DNA for cancer detection (Mutational Analysis Of Plasma DNA For Cancer Detection), which was filed March 13, 3013

At 1050, a plurality of loci are identified to determine the first methylome. These loci may correspond to each of the loci used to determine the cell-free DNA methylation profile and the second methylation profile. Thus, many loci can correspond. Possibly, more loci could be used to determine the cell-free DNA methylation profile and the second methylation profile.

In some embodiments, loci that are hypermethylated or hypomethylated in the second methylation profile can be identified, for example using maternal blood cells. To identify loci that have been hypermethylated in maternal blood cells, one end of the chromosome can be scanned to detect a CpG site with a methylation index >X% (eg, 80%). The next CpG site can then be searched for in the downstream (3') region (eg, downstream 200 bp). If races

- 17 042157 immediately below the CpG site also has a methylation index >X% (or > other specified value), the first and second CpG sites can be grouped. Grouping can continue either until there are no CpG sites left in the next lower region; or until the methylation index of the immediately downstream CpG site is < X%. A region of clustered CpG sites can be described as hypermethylated in maternal blood cells if the region contains at least five directly adjacent hypermethylated CpG sites. A similar analysis to look for loci that were hypomethylated in maternal blood cells can be performed for CpG sites with methylation indices <20%. Methylation densities for a second methylation profile can be calculated for preselected loci and used to establish a first methylation profile (eg, placental tissue methylation density) of the respective loci, eg, from maternal plasma bisulfite sequencing.

At step 1060, the first methylation profile of the first tissue is determined by calculating a differential parameter that includes the difference between the methylation density of the second methylation profile and the methylation density of the cell-free DNA methylation profile for each of the plurality of loci. This difference is expressed as a percentage.

In one embodiment, the first locus methylation density in the first (eg, placental) tissue (D) is derived using the following equation:

Stc-mp) ^D - ^mbc - where mbc is the methylation density of the second methylation profile at the locus (eg, a pre-selected locus determined by bisulfite sequencing in maternal blood cells); mp denotes the methylation density of the corresponding locus in maternal plasma according to bisulfite sequencing; f represents the percentage of cell-free DNA from the first tissue (eg, the fractional concentration of fetal DNA), and CN represents the number of copies at the indicated locus (eg, a higher value for amplification or a lower number for deletions relative to normal). If there are no amplifications or deletions in the first tissue, the CN may be 1. In trisomy (or duplication of an area in a tumor or fetus), the CN may be 1.5 (because there is an increase from 2 copies to 3 copies) and for monosomy it may be equals 0.5. At higher amplification levels, increases in 0.5 steps may occur. In this example, D may correspond to a differential parameter.

In step 1070, the first methylation density is converted to obtain an adjusted first methylation density of the first tissue. This transformation takes into account the fixed differences between the differential parameters and the actual methylation profile of the first tissue. For example, the indicated values may differ by a fixed constant or by a factor. Said transformation may be linear or non-linear.

According to one implementation variant, it was found that the distribution of the derived values, D, is lower than the actual level of placental tissue methylation. For example, inferred values can be linearly transformed using data for CpG islands, which are genomic segments dominated by CpG sites. Genome positions of the CpG islands used in this study were obtained from the UCSC Genome Browser database (NCBI version 36/hg18) (PA Fujita et al. 2011 Nucleic Acids Res; 39: D876-882). For example, a CpG island can be defined as a genomic segment with GC content >50%, genome length >200 bp. and an observed/expected CpG ratio >0.6 (M Gardiner-Garden et al 1987 J Mol Biol; 196:261-282).

In one embodiment, CpG islands with at least 4 CpG sites and an average sequencing depth of >5 per CpG site in sequenced samples can be included to obtain a linear transformation equation. After determining linear relationships between the methylation densities of CpG islands in CVS or mature placenta and the derived values, D, the following equations were used to determine the predicted values.

Predicted values for the first trimester =Dx 1.6+0.2 Predicted values for the third trimester =Dx 1.2+0.05

B. An example with a fetus.

As mentioned above, method 1000 can be used to infer a placental methylation landscape from maternal plasma. The circulating DNA in plasma is mainly derived from hematopoietic cells. However, there is an unspecified proportion of cell-free DNA coming from other internal organs. In addition, placental-derived cell-free DNA constitutes approximately 5-40% of all DNA in maternal plasma, with an average value of approximately 15%. Accordingly, it can be assumed that the level of maternal plasma methylation is equivalent to the sum of the existing background methylation and the contribution of the placenta during pregnancy, as described above.

The maternal plasma methylation level, MR, can be determined using the following equation:

MP = BKG x (1 - L + PLN x / where BKG represents the background level of DNA methylation in plasma derived from blood cells and internal organs, PLN represents the level of placental methylation, and f represents the fractional concentration of fetal DNA in maternal plasma.

According to one implementation variant, the level of methylation of the placenta can theoretically be derived from the equation:

MP-BKG X ¢1-/)

PLN = (2)

Equations (1) and (2) are equivalent if CN is 1, D is PLN, and BKG is mbc. In another embodiment, the fractional concentration of fetal DNA can be considered to take on a specified value, or set at a specified value, for example, under the assumption that a minimum f is present.

Maternal blood methylation was taken as background maternal plasma methylation. In addition to loci that were hypermethylated or hypomethylated in maternal blood cells, the present inventors further explored the clearance pathway by focusing on certain clinically relevant regions, such as CpG islands in the human genome.

Mean methylation density for a total of 27,458 CpG islands (NCBI version 36/hg18) on autosomes and X chromosome was obtained from maternal plasma and placenta sequencing data. Only islets with a coverage of >10 CpG sites and an average sequencing depth of >5 per covered CpG sites were selected in all analyzed samples, including placenta, maternal blood, and maternal plasma. As a result, 26,698 CpG islands (97.2%) were left as eligible, and their methylation level was derived using plasma methylation data and fetal DNA fractional concentration according to the above equation.

It was noted that the distribution of the derived PLN values was below the actual level of methylation of CpG islands in placental tissue. Accordingly, in one embodiment, the deduced PLN values, or simply the deduced values (D), were used as a conventional unit for estimating the level of methylation of CpG islands in the placenta. After the transformation, the derived values were linearized, and their distribution began to more closely resemble the actual data set. The converted derived values were termed methylation predictive values (MPV) and were subsequently used to predict the level of methylation of genetic loci in the placenta.

In this example, CpG islands were classified into 3 categories based on their placental methylation densities: Low (<0.4), Medium (>0.4-<0.8) and High (>0.8). Using the derivation equation, the present inventors calculated the MPV for the same set of CpG islands and then used these values to classify the islets into 3 categories with the same cut-off points. By comparing the actual and inferred datasets, the present inventors found that 75.1% of the pre-selected CpG islands could be correctly matched by MPV to the same tissue data categories. Approximately 22% of CpG islands were assigned to the same group at one level difference (high/medium or medium/low density) and less than 3% were suspected to be completely misclassified (high/low density) (FIG. 12A). The overall classification efficiency was also determined: 86.1, 31.4 and 68.8% of CpG islands with methylation densities <0.4, >0.4-<0.8 and >0.8 in the placenta were correctly considered to have low , medium and high methylation density (Fig. 12B).

In FIG. 11A and B are graphical representations of the performance of a prediction algorithm using maternal plasma data and fetal DNA fractional concentration in accordance with embodiments of the present invention. In FIG. 11A is a graphical representation 1100 showing the classification accuracy of CpG islands using MPV-adjusted classification (the inferred category exactly matches the actual data set); difference by 1 level (the inferred category differs by 1 level from the actual data set); and misclassification (the inferred category is the opposite of the actual dataset). Fig. 11B is a graphical representation 1150 showing the proportion of classifiable CpG islands in each of the derived categories.

Provided that maternal background methylation in the relevant genomic regions is negligible, the presence of hypermethylated placental-derived DNA in the circulation increases the total plasma methylation level to an extent that depends on the fractional concentration of fetal DNA. A pronounced change can be observed when the released fetal DNA is fully methylated. Conversely, if maternal background methylation is significant, the rate of change in plasma methylation will become more significant with the release of hypomethylated fetal DNA. Thus, the application of the specified breeding algorithm may be more appropriate when the level of methylation is derived for genetic loci, which,

- 19 042157 reportedly differ for maternal background and placenta, in particular for hypermethylated and hypomethylated placental markers.

Fig. 12A is a table. 1200 for details of 15 selected genomic loci for predicting methylation according to embodiments of the present invention. To validate the methods, the authors of the present invention selected 15 differently methylated genomic loci that had been investigated previously. The methylation levels of selected regions were derived and compared with the previously studied 15 differently methylated genetic loci (RWK Chiu et al. 2007 Am J Pathol; 170: 941-950; S.S.C. Chim et al. 2008 Clin Chem; 54: 500-511 ; SSC Chim et al. 2005 Proc Natl Acad Sci USA; 102: 14753-14758; DWY Tsui et al. 2010 PLoS One; 5: e15069).

Fig. 12B is a graphical representation of 1250 deduced categories of 15 selected genomic loci and their corresponding placental methylation levels. The following categories of methylation were derived: low, <0.4; mean, >0.4-<0.8; high, >0.8. Tab. 1200 and graphical representation 1300 demonstrate that placental methylation levels can be correctly inferred for them, with a few exceptions: RASSF1A, CGI009, CGI137, and VAPA. Of these 4 markers, only CGI009 showed a noticeable discrepancy with the actual data set. The rest were minimally misclassified.

In table. 1200 1 refers to derived values (D) calculated according to the equation:

where f is the fractional concentration of fetal DNA. Score 2 refers to methylation predictive values (MPV), referring to linearized derived values using the following equation: MPV = D x 1.6 + 0.25. Score 3 refers to classification cut-off points for inferred values: Low, <0.4; Average, >0.4-<0.8; High, >0.8. Score 4 refers to the classification cut-off points for the actual placental dataset: Low, <0.4; Average, >0.4-<0.8; High, >0.8. A score of 5 means placental status refers to the methylation status of the placenta relative to the methylation status of maternal blood cells.

C. Calculation of fractional concentrations of fetal DNA.

In one embodiment, the Y chromosome for a male fetus can be used to determine the percentage of fetal DNA from the first tissue. The proportion of Y chromosome sequences (% chrY) in a maternal plasma sample was a mixture of Y chromosome reads derived from a male fetus and a number of maternal (female) reads that were erroneously aligned to the Y chromosome (RWK Chiu et al. 2011 BMJ; 342 :c7401). Accordingly, the relationship between %chrY and the fractional concentration of fetal DNA (f) in the sample can be given as:

where %chrY _male refers to the ratio of Y-aligned reads in a plasma sample containing 100% male DNA; and %chrY _female refers to the ratio of reads aligned to the Y chromosome in a plasma sample containing 100% female DNA.

The %chrY value can be determined from reads that aligned on the Y chromosome without mismatches from a sample obtained from a female pregnant with a male fetus, for example, from reads obtained from bisulfite-converted samples. The %chrY _male value can be obtained by bisulfite sequencing of two adult male plasma samples. The %chrY _female value can be obtained by bisulfite sequencing of two plasma samples from a non-pregnant adult woman.

In other embodiments, the percentage of fetal DNA can be determined from fetal-specific alleles on the autosome. In another example, epigenetic markers can be used to determine the percentage of fetal DNA. Other methods for determining the percentage of fetal DNA may also be used.

D. Method of using methylation to determine the number of copies.

The placental genome is more hypomethylated than the maternal genome. As discussed above, maternal plasma methylation depends on the fractional concentration of placental-derived fetal DNA in maternal plasma. Thus, by analyzing the methylation density of the chromosomal region, it is possible to detect the difference in the contribution of fetal tissues to maternal plasma. For example, in a pregnant woman carrying a trisomic fetus (for example, having trisomy 21, or trisomy 18, or trisomy 13), said fetus will contribute an additional amount of DNA from the trisomic chromosome to the maternal plasma compared to the disomic chromosomes. In this situation, the plasma methylation density for a trisomic chromosome (or any chromosomal region containing amplification) will be lower than the methylation density for disomic chromosomes. The degree of difference can be predicted by mathematical calculation, taking into account the fractional concentration of fetal DNA in the plasma sample. The higher the fractional concentration of fetal DNA in the plasma sample, the greater the difference in methylation density between

- 20 042157 trisomic and disomic chromosomes. For regions containing a deletion, the methylation density will be higher.

One example of a deletion is Turner syndrome, in which a female fetus has only one copy of the X chromosome. In this situation, a pregnant woman carrying a fetus suffering from Turner syndrome will have a higher density of X-chromosome methylation in plasma DNA than in a situation where the same pregnant woman is carrying a female fetus with a normal number of X chromosomes. In one embodiment of this strategy, maternal plasma may first be analyzed for the presence or absence of Y chromosome sequences (eg, using MPS or a PCR-based technique). If the Y-chromosome sequences are present, the fetus may be classified as a male fetus and no further analysis will be necessary. On the other hand, if there are no Y-chromosome sequences in the maternal plasma, the fetus may be classified as a female fetus. In this situation, an analysis of the X-chromosome methylation density in maternal plasma can be performed. A higher density of X-chromosome methylation compared to normal will indicate a high risk of Turner syndrome in the fetus. This method can also be applied to other sex chromosome aneuploidies. For example, an XYY fetus will have a lower maternal plasma Y-chromosome methylation density than a normal XY fetus with a similar maternal plasma fetal DNA level. In another example, a fetus with Klinefelter syndrome (XXY) has Y-chromosome sequences in maternal plasma, but the maternal plasma X-chromosome methylation density is lower than that of a normal XY fetus with a similar level of fetal DNA in maternal plasma.

Based on the above, the plasma methylation density for the disomic chromosome ίΧΛΡ ϊ A MP _Nm _ _aneu = BKGx(lf) + PLNxf „ (MP _Non _ _aneu ) can be calculated as: > where BKG is the background plasma DNA methylation level obtained from on blood and visceral cells, PLN is the placental methylation level and f is the fractional concentration of fetal DNA in maternal plasma.

The plasma methylation density for a trisomic chromosome (MPAn _eu ) can be calculated „ . MP _Upvi = BKG xQi-f ) + PLN x f xl.5 , .___ _Λ .

as: ^Upwi * > where 1.5 corresponds to the number of copies of CN and the addition of one more chromosome gives an increase of 50%. The difference between trisomic and disomic chromosomes (MP _Diff ) will be

MP _K ff = PLN x / x 0.5

In one embodiment, a comparison of the methylation density of a potentially aneuploid chromosome (or chromosomal region) with one or more other suspected non-aneuploid chromosomes, or a total genome methylation density, can be used to effectively normalize the concentration of fetal DNA in a plasma sample. The comparison can be made by calculating a parameter (eg, including a ratio or difference) relating the methylation densities of the two regions to obtain a normalized methylation density. Said comparison may eliminate the dependent nature of the resulting methylation level (eg, determined as a parameter by two methylation densities).

If the methylation density of a potentially aneuploid chromosome is not normalized to the methylation density of one or more other chromosomes, or to other parameters reflecting the fractional concentration of fetal DNA, the fractional concentration will be the main factor influencing plasma methylation density. For example, the plasma methylation density of chromosome 21 in the plasma of a pregnant woman carrying a fetus with trisomy 21 at a fractional concentration of fetal DNA of 10% will be the same as that of a pregnant woman bearing an euploid fetus at a fractional concentration of fetal DNA of 15%, while the normalized density methylation will show the difference.

In another embodiment, the methylation density of a potentially aneuploid chromosome can be normalized to a fractional concentration of fetal DNA. For example, the following equation can be used to normalize methylation density: Normalized ^-^nan-normalized + (BKG - PLN)xf ' where MP _Normalized is the methylation density normalized to fractional fetal plasma concentration, MP _Non - _normalized is measured methylation density, BKG is the background methylation density in maternal blood cells or tissues, PLN is the methylation density in placental tissues, and 1/2 is the fractional concentration of fetal DNA. The methylation densities of BKG and PLN can be based on reference values previously established from maternal blood cells and placental tissues obtained from healthy pregnancies. Various genetic and epigenetic methods can be used to determine the fractional concentration of fetal DNA in a plasma sample, for example, measuring the percentage of sequencing reads from the Y chromosome using massively parallel sequencing or non-bisulfite-converted PCR

- 21 042157

DNA.

In one embodiment, the normalized methylation density of a potentially aneuploid chromosome can be compared to a reference group, which is a pregnant woman carrying euploid fetuses. The mean and SD of the normalized methylation density of the reference group can be determined. Then, the normalized methylation density in the tested case can be represented as a Z-score reflecting the number of standard Z-scores

SD deviations of the SD from the mean for the reference group, as: ’ where

MPNo.m. zed is the normalized methylation density for the test case, Mean is the mean of the normalized methylation density of the reference cases, and SD is the standard deviation of the normalized methylation density of the reference cases. A cut-off point, such as a Z-score <-3, can be used to classify a chromosome as significantly hypomethylated and thus to determine the sample's aneuploid status.

In another embodiment, MP _Diff can be used as a normalized methylation density. In such an embodiment, the PLN may be derived, for example, using method 1000. In some embodiments, the reference methylation density (which may be normalized using f) may be determined from the methylation level of the non-aneuploid region. For example, the mean (Mean) can be determined from one or more chromosomal regions in one sample. The cut-off point can be scaled to f, or simply set to a level sufficient as long as the minimum concentration is present.

Accordingly, the comparison of the methylation level for the area with the cut-off point can be done in different ways. Said comparison may include normalization (eg, as described above), which may be equivalent to methylation level or cut-off, depending on how the values are determined. Accordingly, it is possible to determine in various ways whether the methylation level of a region is statistically different from the reference level (determined from the same sample or other samples).

The analysis described above can be used in the analysis of chromosomal regions, which may include the entire chromosome or parts of the chromosome, including consecutive or disjointed sub-regions of the chromosome. In one embodiment, a potentially aneuploid chromosome can be divided into a number of regions. The dimensions of these sections may be the same or different. The methylation density of each site can be normalized by the fractional concentration of the sample, or by the methylation density of one or more putatively non-aneuploid chromosomes, or by the total methylation density of the genome. The normalized methylation density of each site can then be compared to the reference group to determine if it has significant hypomethylation. The percentage of significantly hypomethylated regions can then be determined. A cut-off point, such as greater than 5%, 10%, 15%, 20%, or 30% of significantly hypomethylated regions, can be used to classify the test case's aneuploid status.

When testing for amplification or deletion, methylation density may be compared to a reference methylation density, which may be specific to the particular region being tested. Each region may have a different reference methylation density, since methylation may vary from region to region, in particular depending on the size of the regions (eg, smaller regions show more variability).

As mentioned above, one or more pregnant women, each carrying a euploid fetus, can be used to determine the normal range of methylation density of a region of interest, or the difference in methylation density between two chromosomal regions. A normal range may also be determined for PLN (eg, by direct measurement or derived according to method 1000). In other embodiments, the analysis may use the ratio of two methylation densities, eg, a potentially aneuploid chromosome and a non-aneuploid chromosome, instead of their difference. This methylation analysis method can be combined with the sequence read count method (RWK Chiu et al. 2008 Proc Natl Acad Sci USA; 105:20458-20463) and methods involving plasma DNA size analysis (US Pat. No. 2011/0276277) to determine whether or confirmation of aneuploidy. The sequence read count method used in combination with methylation analysis can either be performed using random sequencing (RWK Chiu et al. 2008 Proc Natl Acad Sci USA;105:20458-20463; DW Bianchi DW et al. 2012 Obstet Gynecol 119:890 -901) or directed sequencing (AB Sparks et al. 2012 Am J Obstet Gynecol 206:319.e1-9; B Zimmermann et al. 2012 Prenat Diagn 32:1233-1241; GJ Liao et al. 2012 PLoS One; 7 :e38154).

The use of BKG makes it possible to account for background variations between samples. For example, one woman may have different BKG methylation levels than another woman, but in such a situation, the difference between BKG and PLN for all samples can be used. The cut-off point for different chromosomal regions may be different, for example, if the methylation density of one region of the genome

- 22 042157 differs from the methylation density of another region of the genome.

This method can be generalized to detect any chromosomal aberrations, including deletions and amplifications, in the fetal genome. In addition, the resolution of this analysis can be adjusted to the required level, for example, the genome can be divided into sections of 10 Mb, 5 Mb, 2 Mb, 1 Mb, 500 kb, 100 kb. Accordingly, this technology can also be used to detect subchromosomal duplications or subchromosomal deletions. This technology, accordingly, allows you to establish the prenatal molecular karyotype of the fetus in a non-invasive manner. When applied in this way, this technology can be used in combination with non-invasive prenatal testing methods based on molecular scoring (A Srinivasan et al. 2013 Am J Hum Genet;92:167-176; SCY Yu et al. 2013 PLoS One 8: e60968). In other embodiments, the size of the plots need not be identical. For example, the size of the regions can be adjusted so that each region contains the same number of CpG dinucleotides. In this case, the physical size of the plots will be different.

The equation can be reformulated to apply to different types of chromosomal aberMP _Diff = {BKG - PLN) X / X 0.5 x CN radios as ' where CN is the amount of change in the number of copies in the affected area. CN = 1 for the addition of 1 copy of a chromosome, 2 for the addition of 2 copies of a chromosome, and -1 for the loss of one of the two homologous chromosomes (for example, when fetal Turner syndrome is detected, in which one of the X chromosomes is lost in a female fetus, resulting in to the XO karyotype). The specified equation does not require any edits when changing the size of the parcels. However, sensitivity and specificity may be reduced when smaller regions are used because fewer CpG dinucleotides (or other combinations of nucleotides for which there are differences in fetal and maternal DNA methylation) are present in smaller regions, resulting in an increase in random fluctuations in the measurement of methylation densities. In one embodiment, the required number of reads can be determined by analyzing the methylation density variation coefficient and the required level of sensitivity.

To demonstrate the feasibility of this method, the present inventors examined plasma samples from 9 pregnant women. Each of the five pregnant women carried a euploid fetus, each of the remaining four women carried a fetus with trisomy 21 (T21). Three of the five euploid pregnancies were randomly selected to form a reference group. The remaining two euploid pregnancies (Eu1 and Eu2) and four T21 cases (T21-1, T21-2, T21-3 and T21-4) were investigated using this method to test potential T21 status. Plasma DNA was converted with bisulfite and sequenced using the Illumina HiSeq2000 platform. In one embodiment, the methylation density of individual chromosomes was calculated. Then, the difference in methylation density between chromosome 21 and the average value for the other 21 autosomes was determined to obtain a normalized methylation density (Table 1). The mean and SD of the reference group were used to calculate the Z-score for the six test cases.

Table 1

When using a cut-off point <-3 for the Z-score to classify a specimen as T21, the classification of all cases of euploidy and T21 was correct.

eul Eu2 T21-1 T21-2 T21-3 T21-4 Z-score for MRids between chromosome 21 and other autosomes -1.48 1.09 -4.46 -5.30 -8.06 -5.69

In another embodiment, the genome was divided into 1 Mb regions and the methylation density for each 1 Mb region was determined. The methylation density of all regions on a potentially aneuploid chromosome can be normalized to the median of the methylation density of all regions located on presumably non-aneuploid chromosomes. In one embodiment, for each site, the difference in methylation density from the median of non-aneuploid sites can be calculated. The Z-score can be calculated for the given values using the mean and SD of the reference group. You can determine the percentage of sites where hypomethylation is observed (Table 2) and compare with the percentage corresponding to the cut-off point.

table 2

Using 5% as a cut-off point for significantly more hypomethylated regions on chromosome 21, all cases were classified correctly for T21 status.

eul Eu2 T21-1 T21-2 T21-3 T21-4 Percentage of regions on chromosome 21 with a Z-score for MP^k <-3 0% 0% 33.3% 58.3% 19.4% 52.8%

This method for detecting fetal chromosomal or subchromosomal aberrations based on DNA methylation can be used in combination with methods based on molecular counts, for example, by sequencing (RWK Chiu et al. 2008 Proc Natl Acad Sci USA; 105: 2045820463), or digital PCR ( YMD Lo et al., 2007 Proc Natl Acad Sci USA;104: 13116-13121), or defined

- 23 042157 by changing the size of DNA molecules (US Patent Publication 2011/0276277). Such a combination (e.g., DNA methylation plus molecular count, or DNA methylation plus sizing, or DNA methylation plus molecular count plus sizing) would produce a synergistic effect that would be beneficial in a clinical setting, such as increased sensitivity and/or specificity. For example, the number of DNA molecules that need to be analyzed, for example by sequencing, can be reduced without adversely affecting diagnostic accuracy. This feature allows such tests to be carried out at a lower cost. In another example, for a certain analyzed number of DNA molecules, the combined approach allows the detection of fetal chromosomal or subchromosomal aberrations at a lower fractional concentration of fetal DNA.

In FIG. 13 is a flow diagram of a method 1300 for detecting a chromosomal abnormality from a biological sample of an organism. Said biological sample comprises a cell-free DNA containing a mixture of cell-free DNA originating from a first tissue and from a second tissue. The first tissue may be from a fetus or tumor, and the second tissue may be from a pregnant woman or from a patient.

At step 1310, a plurality of DNA molecules from the specified biological sample are analyzed. Analysis of a DNA molecule may include determining the localization of said DNA molecule within the genome of said organism and determining whether said DNA molecule is methylated at one or more sites. The analysis can be performed by obtaining reads of sequences using methylation-sensitive sequencing, and, accordingly, the analysis can be performed solely on previously obtained DNA data. In other embodiments, the analysis may include actual sequencing or other active data acquisition steps.

Localization determination may include mapping DNA molecules (eg, using sequence reads) to appropriate portions of the human genome, eg, specific regions. According to one implementation variant, if a read is not mapped to an area of interest, it may be ignored.

At 1320, for each of the plurality of sites, the relative amount of DNA molecules that are methylated at that site is determined. In one embodiment, said sites are CpG sites, and may be only certain CpG sites selected using one or more of the criteria mentioned herein. Determining the amount of methylated DNA is equivalent to determining the amount of unmethylated DNA if the normalization is done using the total number of DNA molecules analyzed at a particular site, for example, the total number of sequencing reads.

In step 1330, the first methylation level of the first chromosomal region is calculated based on the respective numbers of DNA molecules methylated at sites within the first chromosomal region. Said first chromosomal region may be of any size, such as those described above. The level of methylation may take into account the total number of DNA molecules aligned with the first chromosomal region, for example, as part of a normalization procedure.

The first chromosomal region may be of any size (eg, an entire chromosome) and may be composed of disparate sub-regions, ie. subregions that are separated from one another. The methylation levels of each subregion can be determined and combined, for example as a mean or median, to determine the methylation level of the first chromosomal region.

In step 1340, the first methylation level is compared to a cut-off value. The cut-off value may represent a reference methylation level or be related to a reference methylation level (eg, be at a certain distance from the normal level). The cut-off value can be determined from other pregnant female subjects bearing fetuses without a chromosomal abnormality in the first chromosomal region, from samples of individuals not suffering from cancer, or from loci of the specified organism that are not known to be associated with aneuploidy (i.e. over areas that are disomic).

In one embodiment, the cut-off value can be defined as ~ xfxO.Sx CN _} away from the reference methylation level, originating from the first tissue, and CN is the number of copies tested. CN is an example of a scaling factor corresponding to the type of anomaly (deletion or duplication). Initially, a CN cutoff of 1 may be used to test all amplifications, and additional cutoffs may be used to determine the extent of amplification. The cut-off value can be based on the fractional concentration of cell-free DNA originating from the first tissue to determine the expected level of methylation for the locus, for example if there is no copy number aberration.

- 24 042157

At step 1350 classify the anomaly of the first chromosomal region based on the specified comparison. A statistically significant difference in levels may indicate an increased risk of a chromosomal abnormality in the fetus. In various embodiments, said chromosomal abnormality may be trisomy 21, trisomy 18, trisomy 13, Turner's syndrome, or Klinefelter's syndrome. Other examples are subchromosomal deletion, subchromosomal duplication or DiGeorge syndrome.

V. Definition of markers

As noted above, some parts of the fetal genome are methylated differently than the maternal genome. These differences may be common to all pregnancies. Regions of differing methylation can be used to identify DNA fragments derived from the fetus.

A. Method for determining DMR from placental tissue and maternal tissue.

The placenta contains tissue-specific methylation signatures. For detection in maternal plasma and for use in non-invasive prenatal diagnosis, fetal-specific DNA methylation markers have been developed based on loci differentially methylated in placental tissues and maternal blood cells (SSC Chim et al. 2008 Clin Chem; 54: 500-511 EA Papageorgiou et al 2009 Am J Pathol; 174: 1609-1618; and T Chu et al. 2011 PLoS One; 6: e14723). Implementations are provided for the analysis of such differentially methylated regions (DMRs) on a genome-wide scale.

Fig. 14 is a flowchart of a method 1400 for identifying methylation markers by comparing the placental methylation profile with the maternal methylation profile (eg, determined from blood cells) in accordance with embodiments of the present invention. The method 1400 can also be used to determine tumor markers by comparing the methylation profile of a tumor with that of healthy tissue.

At step 1410, placental methyl and blood methyl are obtained. Placental bromide can be determined from a placental sample, such as CVS or a mature placenta. It should be understood that methylome may include methylation densities of only a portion of the genome.

At 1420, a region is identified that contains the specified number of sites (eg, 5 CpG sites) and for which a sufficient number of reads have been obtained. In one embodiment, identification begins at one end of each chromosome to locate the first 500 bp region containing at least five relevant CpG sites. A CpG site may be considered suitable if the site is covered by at least five sequence reads.

At 1430, placenta methylation index and blood methylation index are calculated for each site. For example, the methylation index was calculated individually for all eligible CpG sites in each 500 bp region.

In step 1440, the methylation indices of maternal blood cells and the methylation indices of the placental sample were compared to determine if the sets of said indices differ. For example, maternal blood cell methylation indices and CVS or mature placenta methylation indices were compared using, for example, the Mann-Whitney test. A p-value of, for example, <0.01 was considered to indicate a statistically significant difference, although other values could be used, with a lower value resulting in fewer false positive areas.

In one embodiment, if the number of eligible CpG sites was less than 5 or the Mann-Whitney test showed no statistical significance, the 500 bp region shifted lower (in the 3' direction) by 100 p. The region continued to shift in the 3' direction until the Mann-Whitney test showed statistical significance for the 500 bp region. Then the next region of 500 bp in size was considered. If the next region was found to be statistically significant according to the Mann-Whitney test, it was added to the current region, provided that the size of the combined sequential region did not exceed 1000 bp.

At 1450, adjacent regions that have been determined to be statistically significantly different (eg, by the Mann-Whitney test) may be combined. Note that this difference refers to the methylation indices for the two samples. In one embodiment, if adjacent regions are within a specified distance (eg, 1000 bp) from each other and if they exhibit a similar methylation profile, they are pooled. In one embodiment, the similarity of the methylation profile of adjacent regions can be determined based on any of the following: (1) when the same trend is observed in placental tissue relative to maternal blood cells, for example, both regions are more methylated in placental tissues than in blood cells; (2) differences in methylation densities of less than 10% for adjacent areas in placental tissue; and (3) differences in methylation densities of less than 10% for adjacent regions in maternal blood cells.

At step 1460, methylation densities for blood methylome were calculated from cell DNA

- 25 042157 blood and placental sample (eg CVS or mature placental tissue) in the indicated areas. Methylation densities can be determined as described herein.

In step 1470, putative DMRs are determined, characterized in that total placental methylation density and total blood methylation density are statistically significantly different for all sites in the region. In one embodiment, all eligible CpG sites within the pooled region are evaluated using the χ ² test. The ^χ2 test showed whether there was a statistically significant difference between maternal blood cells and placental tissue in terms of the number of methylated cytosines, as a proportion of methylated and unmethylated cytosines in all relevant CpG sites within the pooled region. In one embodiment, for the χ ² test, a P-value of <0.01 may be considered to be indicative of a statistically significant difference. Pooled segments that were shown to be significant using the χ ² test were considered putative DMRs.

In step 1480, loci were identified where maternal blood cell DNA methylation densities were above the upper cutoff or below the lower cutoff. In one embodiment, loci were identified where maternal blood cell DNA methylation densities were either <20% or >80%. In other embodiments, body fluids other than maternal blood may be used, including, but not limited to, saliva, uterine or cervix swabs from a woman's genital tract, tears, sweat, saliva, and urine.

The key to successfully obtaining fetal-specific DNA methylation markers in maternal plasma may be that the methylation status of maternal blood cells is either highly methylated or as unmethylated as possible. This may reduce (eg, minimize) the likelihood of maternal DNA molecules influencing the analysis of placental-derived fetal DNA molecules showing an opposite methylation profile. Accordingly, in one embodiment, candidate DMRs were selected by additional filtering. Loci showing methylation densities <20% in maternal blood cells and at least 20% higher methylation densities in placental tissues were considered candidate hypomethylated loci. Loci showing methylation densities >80% in maternal blood cells and at least 20% lower methylation densities in placental tissues were considered candidate hypermethylated loci. Other percentages may also be used.

Then, at step 1490, DMRs were identified at a subset of loci where placental methylation densities differ significantly from blood methylation densities by comparing said difference to a threshold. In one embodiment, the threshold is 20% such that methylation densities differ by at least 20% from maternal blood cell methylation densities. Accordingly, the difference between placental methylation densities and blood methylation densities at each of the identified loci can be calculated. Said difference may be the result of a simple subtraction. In other implementations, scaling factors and other functions may be applied to determine the difference (eg, the difference may be represented as the result of applying a function to the result of a simple subtraction).

In one embodiment, 11,729 hypermethylated and 239,747 hypomethylated loci were identified using this method from a first trimester placental sample. The first 100 hypermethylated loci are listed in Table 1. S2A applications. The 100 first hypomethylated loci are listed in Table 1. S2B applications. In table. S2A and S2B indicate chromosome, start and end location, region size, maternal blood methylation density, placental sample methylation density, P-values (all P-values are very low), and differences in methylation. Localizations correspond to the hg18 reference genome, which can be found at: hgdownload.soe.ucsc.edu/goldenPath/hg18/chromosomes.

11,920 hypermethylated and 204,768 hypomethylated loci were identified from a third trimester placental sample. The 100 first hypermethylated loci for the third trimester are listed in Table 1. S2C, and the first 100 hypomethylated loci are listed in Table 1. S2D. The 33 loci that were previously described as differentially methylated in maternal blood cells and in first trimester placental tissues were used to validate our list of candidate first trimester loci. Of these 33 loci, 79% were identified as DMR using the algorithm proposed by the authors of the present invention.

Fig. 15A is a table 1500 showing the performance of the DMR identification algorithm when applying first trimester data given the 33 previously described first trimester markers. In the specified table. a means that loci 1-15 have been previously described in (RWK Chiu et al. 2007 Am J Pathol; 170:941-950 and SSC Chim et al. 2008 Clin Chem; 54:500-511); loci 16-23 have been previously described in (KC Yuen, thesis 2007, The Chinese University of Hong Kong, Hong Kong); and loci 24-33 have been previously described in (EA Papageorgiou et al. 2009 Am J Pathol; 174:1609-1618). b means that the specified

- 26 042157 data were obtained from the publications listed above, c means that the methylation densities of maternal blood cells and the chorionic villus sample and their differences were observed in the sequencing data obtained in the present study, but they are based on genomic coordinates given in the original studies, d means that data for loci were identified using implementations of method 1400 on bisulfite sequencing data without using reference data from Chiu et al (2007), Chim et al (2008), Yuen (2007) and Papageorgiou et al (2009) mentioned above . These loci contained previously described genomic regions but tended to cover larger regions, meaning that candidate DMRs were classified as true positive (TP) or false negative (FN) based on meeting the requirement for an observed difference of >0.20 between methylation densities, corresponding to the genomic coordinates of the indicated DMRs in maternal blood cells and in a chorionic villus sample.

Fig. 15B is a table 1550 showing the performance of the DMR identification algorithm using data for the third trimester and comparison with a placental sample obtained at delivery, meaning that the same list of 33 loci was used as in FIG. 17A. b means that because these 33 loci were previously identified from early pregnancy specimens, they may not correspond to data for the third trimester. Therefore, the bisulfite sequencing data obtained in the present study for mature placental tissue based on genomic coordinates obtained in the original studies were revised. A methylation density difference of >0.20 between maternal blood cells and mature placental tissue was used to determine whether these loci were indeed third trimester DMR regions, c means that loci data were identified by applying the 1400 method to bisulfite sequencing data without considering the publications cited above by Chiu et al (2007), Chim et al (2008), Yuen (2007) and Papageorgiou et al (2009). These loci contained the previously described genomic regions but tended to cover larger regions, d means that candidate DMRs containing loci suitable as differentially methylated in the third trimester were classified as true positive (TP) or false negative (FN) on on the basis of compliance with the requirement of an observed difference >0.20 between the methylation densities of the corresponding DMR genomic coordinates in maternal blood cells and in mature placental tissue. For loci not eligible to be differentially methylated in the third trimester, non-listed DMRs or present DMRs containing the indicated loci but showing a methylation difference of <0.20 were considered true negative (TN) DMR regions.

B. DMR based on maternal plasma sequencing.

Placental tissue DMR can be identified directly from maternal plasma DNA bisulfite sequencing, provided that the fractional concentration of fetal DNA in the sample is also known. This is possible because the placenta is the main source of fetal DNA in maternal plasma (SSC Chim et al. 2005 Proc Natl Acad Sci USA 102, 14753-14758) and the present inventors have shown in the present study that the methylation status of fetal-specific DNA in maternal plasma correlates with placental methyloma.

Thus, aspects of method 1400 can be implemented using plasma methylome to clear placental methylome instead of using a placental sample. Accordingly, method 1000 and method 1400 may be combined to determine DMR. Method 1000 can be used to determine placental methylation profile predictors and use them in method 1400. In an example of this assay, attention is also focused on loci that were either <20% or >80% methylated in maternal blood cells.

In one embodiment, to determine loci that were hypermethylated in placental tissues relative to maternal blood cells, the present inventors selected loci that exhibited <20% methylation in maternal blood cells and >60% methylation consistent with a predicted value with a difference of at least 50% between blood cell methylation density and the predicted value. To determine loci that were hypomethylated in placental tissues relative to maternal blood cells, the present inventors selected loci that exhibited >80% methylation in maternal blood cells and <40% methylation as predicted with at least a 50% difference between density methylation of blood cells and the predicted value.

Fig. 16 is a table 1600 showing the number of loci hypermethylated or hypomethylated predicted based on direct analysis of maternal plasma bisulfite sequencing data. H/O means not determined, which means that the search for hypermethylated loci began with a list of loci showing methylation densities <20% in maternal blood cells. b means that the search for hypomethylated loci started with a list of loci showing methylation densities >80% in maternal blood cells, c means that bisulfite sequencing data for a chorionic villus sample was used to validate the data

- 27 042157 maternal plasma of the first trimester, and mature placenta tissue was used to verify the maternal plasma data of the third trimester.

As can be seen from Table. 1600, most of the non-invasively determined loci showed the expected tissue methylation pattern and overlapped with the tissue-based DMRs presented in the section above. The appendix provides a list of DMRs identified in plasma. In table. S3A lists the first 100 loci found to be hypermethylated by bisulfite sequencing for first-trimester maternal plasma. In table. S3B lists the first 100 loci found to be hypomethylated by bisulfite sequencing for first trimester maternal plasma. In table. S3C lists the first 100 loci found to be hypermethylated by bisulfite sequencing for third-trimester maternal plasma. Table S3D lists the first 100 loci found to be hypomethylated by bisulfite sequencing for third-trimester maternal plasma.

C. Gestational variations in placental and fetal methylomes.

The overall proportion of methylated CpGs in CVS was 55%, and 59% for the mature placenta (Table 100 in Figure 1). More hypomethylated DMRs can be identified in CVS than in mature placenta, with some hypermethylated DMRs being similar in both tissues. Accordingly, it was evident that the CVS were more hypomethylated than the mature placenta. This trend in gestation was also evident in the maternal plasma data. The proportion of methylated CpG in fetal-specific reeds was 47.0% in maternal plasma in the first trimester, but 53.3% in maternal plasma in the third trimester. The numbers of validated hypermethylated loci were similar in first trimester (1457 loci) and third trimester (1279 loci) maternal plasma samples; however, first trimester samples had substantially more hypomethylated loci (21812 loci) than third trimester samples (12677 loci) (Table 1600 in Fig. 16).

D. Application of markers.

Differentially methylated markers, or DMRs, are suitable for use in several aspects. The presence of such markers in maternal plasma indicates the presence and confirms the presence of either fetal or placental DNA. This confirmation can be used for quality control in non-invasive prenatal testing. DMRs can serve as universal markers of fetal DNA in maternal plasma and have advantages over markers based on genotypic differences between mother and fetus, such as markers based on polymorphisms or based on the Y chromosome. DMRs are universal fetal markers suitable for all pregnancies. Polymorphism-based markers are only suitable for use in a subset of pregnancies where the fetus inherits said marker from the father and the mother's genome does not contain said marker. In addition, it is possible to measure the concentration of fetal DNA in a maternal plasma sample by quantifying DNA molecules derived from said DMRs.

If there is a known DMR profile expected for normal pregnancies, pregnancy-related complications, in particular involving changes in placental tissue, can be detected by observing deviations in the maternal plasma DMR profile or methylation profile from that expected for normal pregnancies. Pregnancy-related complications involving changes in placental tissue include, but are not limited to, fetal chromosomal aneuploidies. Examples include trisomy 21, preeclampsia, intrauterine growth retardation, and preterm birth.

E. Sets with markers.

In embodiments, compositions and kits for implementing the methods described herein and other suitable methods may be provided. The kits can be used to perform fetal DNA assays, such as cell-free fetal DNA in maternal plasma. In one embodiment, the kit may include at least one oligonucleotide suitable for specific hybridization with one or more of the loci identified herein. The kit may also include at least one oligonucleotide suitable for specific hybridization with one or more reference loci. In one embodiment, placental hypermethylated markers are measured. The test locus may be methylated DNA in maternal plasma and the reference locus may be methylated DNA in maternal plasma. A similar kit can be assembled for the analysis of tumor DNA in plasma.

In some instances, kits may include at least two oligonucleotide primers that can be used to amplify at least a portion of a target locus (eg, the locus provided in the appendix) and a reference locus. Instead of or in addition to primers, the kit may include labeled probes for detecting a DNA fragment corresponding to the target locus and the reference locus. In various embodiments, one or more oligonucleotides of the kit correspond to a locus from the application tables. As a rule, these kits also contain an instruction manual to guide users in the analysis of test samples.

- 28 042157 and assessment of the physiological state or pathology of the test subject.

In various embodiments, a kit is provided for analyzing fetal DNA in a biological sample containing a mixture of fetal DNA and DNA from a subject who is a pregnant woman. The set may include one or more oligonucleotides for specific hybridization with at least a portion of the genomic region listed in tables S2A, S2B, S2C, S2D, S3A, S3B, S3C and S3D. Accordingly, any number of oligonucleotides from any of the tables, or from only one table, may be used. These oligonucleotides can function as primers, and can be presented as primer pairs, with the pair corresponding to a specific region listed in the tables.

VI. Relationship between size and methylation density

It is known that plasma DNA molecules are present in the bloodstream in the form of short molecules, with the majority of molecules being approximately 160 bp in length. (YMD Lo et al. 2010 Sci Transl Med; 2: 61ra91, YW Zheng et al. 2012 Clin Chem; 58: 549-558). Interestingly, the data of the authors of the present invention demonstrate the relationship between the status of methylation and the size of plasma DNA molecules. Accordingly, the length of the plasma DNA fragment is related to the level of DNA methylation. The characteristic size profiles of plasma DNA molecules suggest that most of them are associated with mononucleosomes, possibly formed during enzymatic cleavage during apoptosis.

Circulating DNA in nature is fragmented. In particular, circulating fetal DNA is shorter than maternally derived DNA in maternal plasma samples (KCA Chan et al. 2004 Clin Chem; 50: 88-92). Since the alignment of the paired ends allows analysis of the size of the bisulfite-treated DNA, one can directly assess the presence of any correlation between the size of plasma DNA molecules and their corresponding methylation levels. The authors of the present invention investigated the specified correlation in maternal plasma, as well as in the control sample of the plasma of a non-pregnant adult woman.

Paired-end sequencing (including full molecule sequencing) for both ends of each DNA molecule was used to analyze each sample in this study. By aligning a pair of end sequences of each DNA molecule with the reference human genome and obtaining the genomic coordinates of the ends of the sequencing reads, the lengths of the sequenced DNA molecules can be determined. Plasma DNA molecules are naturally fragmented into small molecules, and plasma DNA sequencing libraries are generally obtained without any fragmentation steps. Accordingly, the lengths determined by sequencing reflect the sizes of the original plasma DNA molecules.

In a previous study, the present inventors determined size profiles of fetal and maternal DNA molecules in maternal plasma (YMD Lo et al. 2010 Sci Transl Med; 2: 61ra91). The authors of the present invention have shown that plasma DNA molecules have dimensions resembling mononucleosomes; fetal DNA molecules were shorter than maternal ones. In this study, the authors of the present invention determined the relationship of the methylation status of plasma DNA molecules with their size.

A. Results.

In FIG. 17A is a graph 1700 showing the size distribution of maternal plasma DNA, non-pregnant control plasma, placental DNA, and peripheral blood DNA. Two bisulfite-treated plasma samples from maternal and non-pregnant female control plasma showed the same characteristic size distributions as previously described (YMD Lo et al. 2010 Sci Transl Med; 2:61ra91), with the most common total sequence length of 166-167 By. and a frequency of 10 b.p. for DNA molecules, the length of which is less than 143 p.

Fig. 17B is a plot 1750 of the size distribution and methylation profile of maternal plasma, adult female control plasma, placental tissue, and adult female control blood. For DNA molecules of the same size and containing at least one CpG site, the average methylation density was calculated. The present inventors then plotted the relationship between the sizes of the DNA molecules and their methylation densities. In particular, the average methylation density was determined for each fragment length ranging from 50 bp to 50 bp. up to 180 bp, for sequencing reads covering at least 1 CpG site. Interestingly, methylation density increased with increasing plasma DNA size, and peaked at approximately 166-167 bp. This pattern, however, was not observed in placental DNA samples and control blood samples fragmented using an ultrasound system.

In FIG. 18 shows plots of methylation densities and plasma DNA molecular size. Fig. 18A is a plot 1800 for maternal plasma in the first trimester. Fig. 18B is a 1850 plot for third trimester maternal plasma. Data for all sequencing reads that covered at least one CpG site are represented by the blue curve 1805. Data for reads that also contained the fetal-specific SNP allele are represented by the red curve 1810.

- 29 042157

Data for reads that also contained the mother-specific SNP allele are represented by the green curve 1815.

Reads containing the fetus-specific SNP allele were considered to belong to fetal DNA molecules. Reads containing the mother-specific SNP allele were considered to belong to maternal DNA molecules. As a rule, DNA molecules with high methylation densities were larger. This trend was observed for molecules of both fetal and maternal DNA, both in the first and third trimesters. As previously mentioned, the overall size of the fetal DNA molecules was smaller than the size of the maternal molecules.

In FIG. 19A is a plot of 1900 methylation densities and sequencing read sizes for an adult non-pregnant woman. For a plasma DNA sample from an adult non-pregnant woman, the same relationship was observed between the size and methylation state of the DNA molecules. On the other hand, during the sonication step prior to MPS analysis, the genomic DNA samples were fragmented. As can be seen in the 1900 plot, there is no trend in the data for blood and placental tissue samples. Since artificial cell fragmentation is performed, it can be expected that there will be no relationship between size and density. Since there is a size dependence for naturally fragmented DNA molecules in plasma, it can be assumed that molecules with lower methylation densities are more likely to break down into shorter fragments.

Fig. 19B is a 1950 graph showing the size distribution and methylation profile of fetal-specific and maternal-specific DNA molecules in maternal plasma. For fetal-specific and maternal-specific plasma DNA molecules, this correlation between fragment sizes and methylation levels was also observed. The lengths of both placental-derived and maternal circulating cell-free DNA fragments increased with increasing methylation levels. In addition, the distributions of their methylation status did not overlap, suggesting that this phenomenon exists regardless of the original fragment length from sources of circulating DNA molecules.

B. Method.

Accordingly, the size distribution can be used to determine the total percent methylation of a plasma sample. This methylation measurement can then be monitored during pregnancy, cancer monitoring, or treatment by serial measurements of plasma DNA size distributions according to the relationship shown in FIG. 18A and B. Measuring methylation can also be used to track increased or decreased release of DNA from an organ or tissue of interest. For example, DNA methylation signatures specific to a particular organ (eg, liver) can be specifically monitored and plasma concentrations of these signatures measured. Since DNA is released into plasma upon cell death, an increase in levels may mean an increase in levels of cell death or damage to that particular organ or tissue. A decrease in the level for a particular organ may mean that the treatment against damage or pathological processes in the specified organ are under control.

Fig. 20 is a flow diagram of a method 2000 for assessing the level of DNA methylation in a biological sample of an organism, in accordance with embodiments of the present invention. The level of methylation can be determined for a specific region of the genome or for the entire genome. If a particular region is desired, DNA fragments exclusively from that particular region may be used.

In step 2010, the numbers of DNA fragments corresponding to different sizes are measured. For each size of a plurality of sizes, the number of a plurality of DNA fragments from a specified biological sample corresponding to a specified size can be measured. For example, the number of DNA fragments with a length of 140 bases can be measured. The indicated amounts can be saved as a bar graph. In one embodiment, the size of each of a plurality of nucleic acids from said biological sample is measured, which can be done on an individual basis (eg, by unimolecular sequencing of the whole molecule or only the ends of the molecule) or on a group basis (eg, by electrophoresis). The sizes can be according to the size range. Accordingly, the amount may refer to DNA fragments having a size falling within a certain range. When conducting sequencing of paired ends, mapping (alignment) of DNA fragments (determined by paired sequence reads) to a specific region can be used to determine the level of methylation of the specified region.

In step 2020, the first value of the first parameter is calculated based on the number of DNA fragments of different sizes. In one aspect, the first parameter provides a statistical measure of the size profile (eg, a histogram) of DNA fragments in a biological sample. This parameter may be referred to as a size parameter since it is determined based on the sizes of a plurality of DNA fragments.

- 30 042157

The first parameter can take various forms. One parameter is the percentage of DNA fragments of a particular or falling within a certain size range relative to all DNA fragments, or relative to DNA fragments of a different size or range. Such a parameter is the number of DNA fragments of a particular size divided by the total number of fragments, which can be obtained from a histogram (any data structure that allows you to count the absolute or relative number of fragments of a particular size). In another example, the parameter may be the number of fragments of a particular size or range, divided by the number of fragments of a different size or range. The division can serve as a normalization to take into account the different number of analyzed DNA fragments for different samples. Normalization can be done by analyzing the same number of DNA fragments for each sample, effectively providing the same result as dividing by the total number of fragments analyzed. Additional examples of parameters and dimensional analysis can be found in US Patent Application 13/789553, which is incorporated herein by reference in all respects.

In step 2030, the first size value is compared with the reference size value. The reference size value can be calculated from the DNA fragments of the reference sample. To determine reference size values, the methylation profile for the reference sample can be calculated and quantified, as well as the value of the first size parameter. Accordingly, by comparing the first size value and the reference size value, the level of methylation can be determined.

At step 2040 determine the level of methylation based on the specified comparison. In one embodiment, it is possible to determine whether the first value of the first parameter is higher or lower than the size reference value, and thus determine whether the methylation level of the test sample is higher or lower than the methylation level for the size reference value. According to another implementation variant, the comparison is carried out by introducing the first value into the calibration function. The calibration function allows you to effectively compare the first value with the calibration values (a set of size reference values) by identifying the point on the curve corresponding to the first value. The calculated methylation level is then obtained as the output value of the calibration function.

Accordingly, it is possible to calibrate the size parameter by the level of methylation. For example, the level of methylation can be measured and associated with a specific dimensional parameter for the specified sample. The data points for different samples can then be approximated by a calibration function. In one embodiment, different calibration functions may be applied to different DNA subsets. Accordingly, some form of calibration may be applied based on prior knowledge of the relationship between methylation and size for a particular DNA subset. For example, the calibration for fetal and maternal DNA may be different.

As shown above, the placenta is more hypomethylated compared to maternal blood, and, accordingly, fetal DNA is smaller due to the lower degree of methylation. Accordingly, the average sample fragment size (or other statistic) can be used to determine the methylation density. Since fragment sizes can be measured using paired-end sequencing rather than the potentially more sophisticated technically sensitive methylation sequencing, this method may be potentially more economical in clinical use. This method can be used to monitor methylation changes associated with the progress of pregnancy or pregnancy-related disorders such as preeclampsia, preterm birth and fetal disorders (eg caused by chromosomal or genetic abnormalities or intrauterine growth retardation).

According to another implementation variant of the specified method can be used for the detection and monitoring of cancer. For example, upon successful treatment of a cancer, the methylation profile in plasma or other body fluid, as assessed using this size-based method, will change towards the profile of healthy individuals who do not have cancer. Conversely, if the cancer progresses, the methylation profile in plasma or other body fluid will diverge from that of healthy individuals who do not have cancer.

In general, hypomethylated plasma molecules were shorter than hypermethylated ones. This trend was observed for both fetal and maternal DNA molecules. Since DNA methylation is known to affect nucleosome packing, the present inventors' data suggest that it is possible that hypomethylated DNA molecules were less densely packed with histones and were thus more susceptible to enzymatic degradation. On the other hand, the data presented in Fig. 18A and 18B also show that despite significantly more fetal DNA hypomethylation compared to maternal reads, the size distributions of fetal and maternal DNA are not completely distinct from each other. In FIG. 19B, it can be seen that even within the same size category, the methylation levels of fetal-specific and maternal-specific species differ.

- 31 042157 apart. This observation suggests that the hypomethylated state of fetal DNA is not the only factor that determines its relatively short length compared to maternal DNA.

VII. Imprinting status of gene loci

Fetal-derived DNA molecules having the same genotype as the mother but different epigenetic signatures can be detected in maternal plasma (LLM Poon et al. 2002 Clin Chem; 48: 35-41). To demonstrate that the sequencing method is a sensitive method for selecting fetal-derived DNA molecules in maternal plasma, the present inventors used the same strategy to detect imprinted fetal alleles in a maternal plasma sample. Two genomic imprinting regions were identified: H19 (chr11:1 977 419-1 977 821, NCBI version 36/hg18) and MEST (chr7:129 917 976-129 920 347, NCBI version 36/hg18). Both contain informative SNPs to distinguish between maternal and fetal sequences. For H19, a maternally expressed gene, the mother was homozygous (A/A) and the fetus was heterozygous (A/C) for SNP rs2071094 (chr11:1 977 740) in this region. One of the maternal A alleles was fully methylated and the other was unmethylated. In the placenta, the A allele was not methylated, while the C allele inherited from the father was fully methylated. The present inventors have detected two methylated C genotype reads corresponding to imprinted placenta-derived paternal alleles in maternal plasma.

MEST, also known as PEG1, is a paternally expressed gene. Both mother and fetus were heterozygous (A/G) for the SNP rs2301335 (chr7:129 920 062) at the specified imprinting locus. In maternal blood, the G allele was methylated and the A allele was unmethylated. The methylation pattern in the placenta was reversed and included a methylated maternal A allele and an unmethylated paternal G allele. In maternal plasma, three unmethylated G alleles could be detected that were paternally derived. In contrast, for VAV1, for the non-imprinting gene locus on chromosome 19 (chr19:6 723 621-6 724 121), no allelic DNA methylation pattern was observed in tissue and plasma samples.

Thus, methylation status can be used to determine which DNA fragments belong to the fetus. For example, maternal plasma A allele detection alone cannot be used as a fetal marker if the mother is heterozygous (GA). However, if there is a distinct plasma A methylation status, methylated A molecules are fetal specific and unmethylated A molecules are maternal specific, or vice versa.

Further, the present inventors have focused on loci that have been described as showing genomic imprinting in placental tissues. Based on the list of loci provided by Woodfine et al. (2011 Epigenetics Chromatin; 4:1), the authors of the present invention additionally selected those loci that contained SNPs in the imprinting control region. Four loci met the specified criteria; these loci were represented by H19, KCNQ10T1, MEST and NESP.

As for the reads from the maternal blood cell sample for H19 and KCNQ10T1, the maternal reads were homozygous for SNPs and the proportions of methylated and unmethylated reads were approximately equal. CVS and mature placental tissue samples showed that the fetus was heterozygous for both loci, and each allele was either strictly methylated or unmethylated, i.e. monoallelic methylation was observed. In maternal plasma samples, fetal DNA molecules inherited from the father were found for both loci. For H19, paternally inherited molecules were represented by sequencing reads that contained a fetal-specific allele and were methylated. For KCNQ10T1, paternally inherited molecules were represented by sequencing reads that contained a fetal-specific allele and were not methylated.

On the other hand, the mother was heterozygous for both MEST and NESP. In the case of MEST, both mother and fetus were heterozygous (GA) for SNPs. At the same time, as is obvious from the data for the 5'-3' chain for maternal blood cells and placental tissue, the methylation status for CpG adjacent to SNPs was opposite in mother and fetus. Allele A was unmethylated in maternal DNA but methylated in fetal DNA. In the case of MEST, the maternal allele was methylated. Accordingly, it can be accurately determined that the fetus has inherited the A allele from the mother (methylated in CVS) and the mother has inherited the A allele from her father (unmethylated in maternal blood cells). Interestingly, in maternal plasma samples, all four groups of molecules can be easily distinguished, including each of the two maternal alleles and each of the two fetal alleles. Thus, by combining genotypic information with methylation status at imprinting loci, the present inventors could easily distinguish maternally inherited fetal DNA molecules from background maternal DNA molecules (LLM Poon et al. 2002 Clin Chem; 48: 35-41).

This method can be used to detect uniparental disomy. For example, if the father of a given fetus is known to be homozygous for the G allele, the failure to detect the unmethylated G allele in maternal plasma means there is no contribution from the paternal allele. In addition, under such circumstances, if both the methylated G allele and the methylated A allele were detected in the plasma of the indicated pregnancy, the fetus is presumably maternally heterogeneous.

- 32 042157 rhodisomy, i.e. inheritance of two different alleles from the mother without inheritance from the father. Alternatively, if both the methylated A allele (the fetal allele inherited from the mother) and the unmethylated A allele (the maternal allele inherited by the mother from her father) were detected in maternal plasma without the unmethylated G allele (the paternal allele that should have been inherited by the fetus) , the fetus presumably has maternal isodisomy, i.e. inheritance of two identical alleles from the mother without inheritance from the father.

In the case of NESP, the mother was heterozygous (GA) for the SNP and the fetus was homozygous for the G allele. The paternal allele was methylated for NESP. In maternal plasma samples, paternally inherited fetal G alleles that were methylated can be easily distinguished from background maternal G alleles that were unmethylated.

VIII. Cancers/donors

Some implementations can be used for detection, screening, monitoring (eg, relapse, remission, or response (eg, presence or absence of response) to treatment), staging, classification (eg, to help select the most appropriate treatment), and cancer prediction. using a circulating plasma/serum DNA methylation assay.

Aberrant methylation is known to occur in cancer DNA (JG Herman et al. 2003 N Engl J Med; 349: 2042-2054). For example, CpG island promoters of genes, such as tumor suppressor genes, are hypermethylated, while CpG sites in the gene body are hypomethylated compared to non-cancer cells. Provided that the methylation profile of cancer cells can be represented by the methylation profile of tumor-derived plasma DNA molecules using the methods described herein, the authors of the present invention assume that the overall plasma methylation profile will be different in individuals with cancer, compared with healthy individuals who do not have cancer, or compared with individuals who are cured of cancer. Types of methylation profile differences can refer to quantitative differences in genome methylation densities and/or methylation densities of segments of genomes. For example, in the plasma of cancer patients, there will be a decrease in the methylation densities of plasma methylome or segments of the genome due to the general hypomethylated nature of cancer tissue DNA (Gama-Sosa MA et al. 1983 Nucleic Acids Res; 11: 6883-6894).

Qualitative changes in the methylation profile should also be reflected in the plasma methylome data. For example, plasma DNA molecules derived from genes that are hypermethylated exclusively in cancer cells will show hypermethylation in the plasma of a cancer patient relative to plasma DNA molecules derived from the same genes but present in a healthy control sample. Because aberrant methylation occurs in most cancers, the methods described herein can be used to detect all forms of aberrant methylation cancers, such as, but not limited to, cancers in the lung, breast, colon and rectum, prostate, nasopharynx, stomach, testicles, skin, nervous system, bones, ovary, liver, hematological tissues, pancreas, uterus, kidneys, bladder, lymphoid tissues, etc. Malignancies can be represented by various histological subtypes such as carcinomas, adenocarcinomas, sarcomas, fibroadenocarcinomas, neuroendocrine and undifferentiated, etc.

On the other hand, the present inventors suggest that tumor-derived DNA molecules can be distinguished from background non-tumor-derived DNA molecules, since the overall small size profile of tumor-derived DNA is more pronounced in DNA molecules derived from loci with tumor-associated aberrant hypomethylation, which should have an additional effect on the size of the DNA molecule. Also, tumor-derived plasma DNA molecules can be distinguished from background non-tumor-derived plasma DNA molecules using several features associated with tumor DNA, including, but not limited to, single nucleotide variations, copy number increase and decrease, translocation, inversion, aberrant hyper- or hypomethylation and size profile. Since all of these changes can occur independently, the combined use of these features may provide an additional advantage for the sensitive and specific detection of cancer DNA in plasma.

A. Size and cancers.

The size of tumor-derived DNA molecules in plasma also resembles the size of mononucleosomal units; these molecules are shorter than the background non-tumor-derived DNA molecules that are also present in the plasma of cancer patients. Dimensional parameters have been shown to correlate with cancers as described in US Patent Application 13/789553, incorporated herein by reference in all respects.

Because a relationship has been demonstrated between size and methylation status of the molecule

- 33 042157 for both fetal and maternal DNA in plasma, the same trend is expected for tumor-derived DNA molecules. For example, hypomethylated molecules will be shorter than hypermethylated molecules in the plasma of cancer patients or subjects being screened for cancer.

B. Methylation densities of different tissues in cancer patients.

In this example, the present inventors examined plasma and tissue samples from a patient with hepatocellular carcinoma (HCC). Blood samples were obtained from the indicated patient with HCC before surgical removal of the tumor and 1 week after surgical removal of the tumor. Plasma and buffy coat were collected after centrifugation of blood samples. The resected tumor and adjacent non-tumor liver tissue were harvested. DNA samples extracted from plasma and tissue samples were analyzed using massively parallel sequencing with and without bisulfite pretreatment. Also investigated the plasma DNA of four healthy individuals without cancer as controls. Bisulfite treatment of DNA converts unmethylated cytosine residues to uracil. During subsequent PCR and sequencing, these uracil residues will behave like thymidine. On the other hand, treatment with bisulfite does not result in the conversion of methylated cytosine residues to uracil. After massively parallel sequencing, sequence reads were analyzed using Methy-Pipe (P Jiang, et al. Methy-Pipe: An integrated bioinformatics data analysis pipeline for whole genome methylome analysis). , paper presented at the International Institute of Electrical and Electronics Engineers (IEEE International Conference on Bioinformatics and Biomedicine Workshops), Hong Kong, December 18-21, 2010, to determine the methylation status of cytosine residues at all positions with dinucleotides CG, i.e. in CpG sites.

Fig. 21A is a table. 2100, which shows the methylation densities of preoperative plasma and tissue samples from a patient with HCC. CpG methylation density for regions of interest (eg, CpG sites, promoter or repeat regions, etc.) refers to the proportion of reads showing CpG methylation in the total number of reads covering genomic CpG dinucleotides. The methylation densities of the buffy coat and non-tumor liver tissue are similar. The total methylation density of tumor tissue, based on data for all autosomes, was 25% lower than the methylation density of buffy coat and non-tumor liver tissue. Hypomethylation was stable for each individual chromosome. The density of plasma methylation was between the density values for non-malignant tissues and cancerous tissues. This observation is consistent with the fact that both cancerous and non-cancerous tissues contribute to the circulating DNA of a cancer patient. The hematopoietic system has been shown to be the main source of circulating DNA in individuals without active malignant conditions (YYN Lui, et al. 2002 Clin Chem; 48: 421-7). Therefore, the authors of the present invention also examined plasma samples obtained from four healthy controls. The number of sequencing reads and the achieved sequencing depth per sample are shown in Table 2150 in FIG. 21B.

Fig. 22 is a table. 220 for autosomal methylation densities ranging from 71.2% to 72.5% in plasma samples from healthy controls. These data indicate the expected level of DNA methylation in plasma samples obtained from individuals lacking a source of tumor DNA. In a cancer patient, the tumor tissue also releases DNA into the bloodstream (KCA Chan et al. 2013 Clin Chem; 59: 211-224); RJ Leary et al. 2012 Sci Transl Med; 4:162ra154). Due to the hypomethylated nature of the HCC tumor, the presence of both tumor-derived and non-tumor-derived DNA in the patient's preoperative plasma resulted in a decrease in methylation density relative to plasma levels of healthy controls. In fact, the methylation density in the preoperative plasma sample was between the values of the methylation densities of the tumor tissue and the plasma of healthy controls. The reason is that the plasma DNA methylation level of cancer patients is affected by the degree of aberrant methylation, in this case hypomethylation, of the tumor tissue and the fractional concentration of tumor-derived DNA in the circulation. A lower methylation density of tumor tissue and a higher fractional concentration of tumor-derived DNA in the bloodstream result in a lower plasma DNA methylation density in a cancer patient. Most tumors are reported to show global hypomethylation (JG Herman et al. 2003 N Engl J Med; 349: 2042-2054; MA Gama-Sosa et al. 1983 Nucleic Acids Res; 11: 6883-6894). Accordingly, the observational data for HCC samples should also be applicable to other types of tumors.

In one embodiment, plasma DNA methylation density can be used to determine the fractional concentration of tumor-derived DNA in a plasma/serum sample if the methylation level of the tumor tissue is known. The level of methylation, such as the methylation density of the tumor tissue, can be determined if a tumor sample or tumor biopsy is available. According to another implementation variant, information regarding the level

- 34 042157 tumor tissue methylation can be obtained by studying the level of methylation in a group of tumors of the same type, and the specified information (for example, the average value of the level or median level) is used for the studied patient using the technology described in the specified invention. The level of methylation of the tumor tissue can be determined by analysis of the patient's tumor tissue or inferred from the analysis of tumor tissues of other patients with the same or similar type of cancer. Tumor tissue methylation can be determined using a variety of methylation-sensitive platforms, including, but not limited to, massively parallel sequencing, single molecule sequencing, microarrays (e.g., oligonucleotide arrays), or mass spectrometry (e.g., Epityper assay from Sequenom Inc. ). In some embodiments, such analysis may be preceded by procedures sensitive to the methylation status of DNA molecules, including, but not limited to, cytosine immunoprecipitation and cleavage with methylation-sensitive restriction enzymes. If the level of tumor methylation is known, after analysis of plasma methylome, the fractional concentration of tumor DNA in the plasma of cancer patients can be calculated.

The relationship between plasma methylation level, P, with tumor DNA fractional concentration, f, and tumor tissue methylation level, TUM, can be described as ΒΚθ ^X (1J)+TUM where BKG is the background DNA methylation level in plasma derived from cells blood and other internal organs. For example, it was shown that the total methylation density of all autosomes is 42.9% in a tumor tissue biopsy obtained from the specified patient with HCC, i.e. this is the TUM value for the specified case. The mean value of the methylation density of plasma samples from four healthy controls was 71.6%, i.e. the BKG value for the specified case. The plasma methylation density for preoperative plasma was 59.7%. Using the indicated values, it is determined that/is 41.5%.

In another embodiment, the methylation level of tumor tissue can be determined non-invasively from plasma methylome data if the fractional concentration of tumor-derived DNA in the plasma sample is known. The fractional concentration of tumor-derived DNA in a plasma sample can be determined by other genetic analysis, such as genome-wide allele loss analysis (GAAL) and single nucleotide mutation analysis as previously described (US Patent Application 13/308473; KCA Chan et al. 2013 Clin Chem; 59: 211-24). This calculation is based on the same relationships as described above, except that in this embodiment, the value of f is known and the value of TUM becomes unknown. Said derivation can be carried out for the entire genome or for parts of the genome, similar to the data discussed in the context of determining the level of placental tissue methylation from maternal plasma data.

In another embodiment, variability between sites or methylation density profiles can be used to distinguish between cancer and non-cancer subjects. The resolution of the methylation assay can be further increased by subdividing the genome into regions of a specific size, eg 1 Mb. In such an embodiment, the methylation density of each 1 Mb patch was calculated for collected samples, e.g., buffy coat, resected HCC tissue, non-tumor liver tissue adjacent to the tumor, and plasma collected before and after tumor removal. According to another implementation variant, the sizes of the plots do not have to be constant. In one embodiment, the number of CpG sites remains constant within each region, while the size of the region itself may vary.

In FIG. 23A and B show the methylation density of buffy coat, tumor tissue, non-tumor liver tissue, preoperative plasma, and postoperative plasma from a patient with HCC. Fig. 23A is a plot 2300 of results for chromosome 1. FIG. 23B is a plot of 2350 results for chromosome 2.

For most 1 Mb sites, the methylation densities of the buffy coat and non-tumor liver tissue adjacent to the tumor were similar, while the methylation densities of tumor tissues were lower. Preoperative plasma methylation density values are between those for tumor and non-malignant tissues. The methylation densities of the studied genomic regions in tumor tissues can be derived using data on preoperative plasma methylation and the fractional concentration of tumor DNA. This method is identical to the method described above using methylation density values for all autosomes. The derivation of tumor methylation values described can also be carried out using said higher resolution data on plasma DNA methylation. Other sizes may also be used, such as 300 kb, 500 kb, 2 Mb, 3 Mb, 5 Mb, or more than 5 Mb. According to one implementation variant, the sizes of the plots do not have to be constant. According to one embodiment, the number of CpG sites remains constant in

- 35 042157 composition of each section, while the size of the section itself may vary.

C. Comparison of methylation density in the plasma of a cancer patient and healthy individuals.

As can be seen from 2100, DNA methylation densities of preoperative plasma in a cancer patient were lower than methylation densities of non-malignant tissues. This is presumably due to the presence of DNA from tumor tissue that is hypomethylated. This lower plasma DNA methylation density can potentially be used as a biomarker for cancer detection and monitoring. When monitoring cancer, if the cancer is progressing, there will be an increased amount of cancer-derived DNA in plasma over time. According to this example, an increased amount of circulating cancer-derived DNA in plasma will lead to a subsequent decrease in plasma DNA methylation density on a genome-wide scale.

Conversely, if the cancer responds to treatment, the amount of cancer-derived DNA in plasma will decrease over time. According to this example, a decrease in the amount of cancer-derived DNA in plasma will lead to an increase in the density of DNA methylation in plasma. For example, if a lung cancer patient with an epidermal growth factor receptor mutation has been treated with a targeted therapy such as tyrosine kinase inhibition, an increase in plasma DNA methylation density is indicative of a response. The subsequent emergence of a tumor clone resistant to tyrosine kinase inhibitors will be associated with a decrease in plasma DNA methylation density, which may indicate relapse.

Plasma methylation density measurements can be performed serially; with such measurements rates of change can be calculated and used to predict or correlate with clinical progression or remission or prognosis. For selected genomic loci that are hypermethylated in cancerous tissues but hypomethylated in normal tissues, such as the promoter regions of certain tumor suppressor genes, the relationship between cancer progression and favorable response to treatment will be inverse to the patterns described above.

To demonstrate the feasibility of this method, the present inventors compared the DNA methylation densities of plasma samples obtained from a cancer patient before and after surgical removal of the tumor with plasma DNA obtained from four healthy controls.

Table 2200 lists the DNA methylation densities of each autosome and the pooled values for all autosomes of preoperative and postoperative cancer patient plasma samples, and the pooled values for all autosomes of four healthy controls. For all chromosomes, the DNA methylation densities in the preoperative plasma sample were lower than those in the postoperative and plasma samples from four healthy subjects. The difference in plasma DNA methylation densities between preoperative and postoperative samples provides evidence that the lower methylation densities in the preoperative plasma sample were due to the presence of DNA from the HCC tumor.

The return of DNA methylation densities in the postoperative plasma sample to levels similar to those in plasma samples of healthy controls suggests that a significant portion of the tumor-derived DNA has disappeared due to surgical removal of the source, i.e. tumors. These data suggest that preoperative plasma methylation density, as determined using data available for large genomic regions such as all autosomes or individual chromosomes, was at a lower level than healthy controls, thus identifying the test case as a cancer. those. establish a diagnosis or screening.

Preoperative plasma data also showed significantly lower methylation levels than postoperative plasma, indicating that plasma methylation levels can also be used to monitor tumor burden and thus predict and monitor patient cancer progression. Reference values can be determined from the plasma of healthy controls or individuals who are at risk of developing cancer but are currently cancer free. Individuals at risk of developing HCC include those with chronic hepatitis B or hepatitis C infection, individuals with hemochromatosis, and individuals with cirrhosis of the liver.

Plasma methylation density values beyond, for example below, a specified cut-off point based on reference values can be used to assess the presence or absence of tumor DNA in a non-pregnant individual's plasma. For detection of the presence of hypomethylated circulating tumor DNA, the cut-off point can be defined as being below the 5th or 1st percentile of values in the control population, or can be defined based on several standard deviations, such as 2 or 3 standard deviations (SD), below the mean values. methylation density in controls, or based on multiple median values (MoM). For hypermethylated tumor DNA, the cut-off point can be defined as a level above 95

- 36 042157 or the 99th percentile of the values in the control population, or can be determined based on several standard deviations, for example, 2 or 3 SD in the mean values of the methylation density of the values of the controls, or based on the determination of several median values (MoM). In one embodiment, the age in the control population matches the age of the test subject. Age matching does not have to be exact and can be represented by age groups (eg, 30 to 40 years old, with the test subject being 35 years old).

The present inventors then compared the methylation densities of 1 Mb patches in plasma samples from a cancer patient and in samples from four control subjects. For clarity, the indicated results for chromosome 1 are presented.

Fig. 24A is a graph 2400 depicting preoperative plasma methylation densities from a patient with HCC. Fig. 24B is a graph 2450 depicting postoperative plasma methylation densities from a patient with HCC. Blue dots represent results for control subjects, red dots represent results for a plasma sample from a patient with HCC.

As seen in FIG. 24A, methylation densities in preoperative plasma from the HCC patient were lower than methylation densities in control subjects for most sites. Similar patterns were observed for other chromosomes. As seen in FIG. 24B, methylation densities in postoperative plasma from the HCC patient were similar to methylation densities in control subjects for most sites. Similar patterns were observed for other chromosomes.

To assess whether a test subject has cancer, the result for the test subject is compared with the values for the reference group. According to one implementation variant, the reference group can be composed of several healthy subjects. In another embodiment, the reference group may be composed of subjects with non-malignant conditions, such as chronic hepatitis B infection or cirrhosis. The difference in methylation densities between the test subject and the reference group can then be quantified.

According to one implementation variant, the reference range can be determined from the values of the control group. The deviations of the result for the test subject from the upper or lower limit for the reference group can then be used to determine whether the subject has a tumor. This value will be influenced by the fractional concentration of tumor-derived DNA in plasma and the difference in methylation levels between malignant and non-malignant tissues. The higher fractional concentration of tumor-derived DNA in plasma results in large differences in methylation density between the tested plasma sample and controls. The large degree of difference in methylation levels between malignant and non-malignant tissues is also associated with large differences in methylation density between the tested plasma sample and controls. In yet another embodiment, different reference groups are selected for test subjects from different age ranges.

In another embodiment, the mean and SD of the methylation densities of four control subjects were calculated for each 1 Mb region. The difference between the methylation densities of the HCC patient and the mean for control subjects was then calculated for the respective sites. In one embodiment, said difference is further divided by the SD of the respective area to determine the Z-score. In other words, the Z-score represents the difference in the methylation densities of test and control plasma samples, expressed as several SD means for control subjects. A Z-score >3 for a region means that HCC patient plasma DNA is more hypermethylated than controls by more than 3 SD at that region, while a <-3 Z-score for a region means that HCC patient plasma DNA more hypomethylated than control subjects, more than 3 SD in the specified area.

In FIG. 25A and B are Z-scores of plasma DNA methylation densities for preoperative (plot 2500) and postoperative (plot 2550) plasma samples from an HCC patient using plasma methylome data from four healthy controls as a reference for chromosome 1. Each point represents a result for one section of 1 MB. Black dots indicate areas with a Z-score ranging from -3 to 3. Red dots indicate areas with a Z-score <-3.

Fig. 26A is a table. 2600 for preoperative and postoperative plasma Z-score data. Most of the sites on chromosome 1 (80.9%) in the preoperative plasma sample had a Z-score of <-3, indicating that the HCC patient's preoperative plasma DNA was significantly more hypomethylated compared to controls. On the other hand, the number of red dots was significantly reduced in the postoperative plasma sample (8.3% of the sites on chromosome 1), suggesting that most of the tumor DNA was eliminated from the circulation due to surgical removal of the source of circulating tumor DNA.

Fig. 26B is a Circos plot of the 2620 showing the Z-score of the densities of

- 37 042157 plasma DNA tylation for preoperative and postoperative HCC patient plasma samples using four healthy controls as a reference for 1 Mb regions analyzed for all autosomes. The outer circle represents the ideograms of human autosomes. The middle circle represents the data for the preoperative plasma sample. The inner circle represents the indicated data for the postoperative plasma sample. Each dot represents the result for one 1 MB chunk. Black dots indicate areas with Z-scores from -3 to 3. Red dots indicate areas with Z-scores <-3. Green dots indicate areas with Z-scores >3.

Fig. 26C is a table 2640 showing the whole genome Z-score distribution of 1 Mb regions in both preoperative and postoperative HCC patient plasma samples. These results indicate that HCC patient preoperative plasma DNA was more hypomethylated than controls for most regions (85.2% of 1 Mb regions) across the entire genome. On the other hand, the majority of areas (93.5% of 1 Mb patches) in the postoperative plasma sample did not show significant hypermethylation or hypomethylation compared to controls. These data indicate that most of the tumor DNA, inherently mostly hypomethylated for this HCC, was no longer present in the postoperative plasma sample.

In one embodiment, the number, percentage, or proportion of sites with Z-scores <-3 may be used to determine the presence of cancer. For example, as shown in Table 2640, 2330 of 2734 sites analyzed (85.2%) had Z-scores <-3 in preoperative plasma, but only 171 of 2734 sites analyzed (6.3%) had Z-scores <-3 in postoperative plasma. These data indicate that the tumor DNA load in preoperative plasma was significantly higher than in postoperative plasma.

Limit values for the number of sites can be determined using statistical methods. For example, approximately 0.15% of parcels are expected to have a Z-score <-3 based on a normal distribution. Thus, the cut-off point for the number of sites may be 0.15% of the total number of sites analyzed. In other words, if more than 0.15% of regions with Z-scores <-3 are observed in a plasma sample of a non-pregnant individual, there is a source of hypomethylated DNA in plasma, namely cancer. For example, 0.15% of the 2734 1 Mb patches that the present inventors examined in this example is approximately 4 patches. Using this value as the cut-off point, both preoperative and postoperative plasma samples contained hypomethylated tumor-derived DNA, although the amount was significantly higher in the preoperative plasma sample compared to the postoperative plasma sample. In four healthy control subjects, none of the sites showed significant hypermethylation or hypomethylation. Other limits may apply (eg 1.1%) and may vary depending on the requirements of the assay used. In other examples, the cut-off percentage may vary based on statistical distribution as well as desired sensitivity and acceptable specificity.

In another embodiment, the cut-off point can be determined by analyzing performance curves (ROC curves), by analyzing the number of patients with cancer and cancer-free individuals. To further validate the specificity of this approach, a plasma sample from a patient seeking medical advice for a non-malignant condition (C06) was examined. 1.1% of plots had a Z-score <-3. In one embodiment, different thresholds may be used to classify different levels of disease status. A lower threshold percentage can be used to distinguish health status from benign conditions, and a higher threshold percentage to distinguish between benign conditions and malignancies.

The diagnostic performance of plasma hypomethylation analysis using massively parallel sequencing appears to be superior to that of polymerase chain reaction (PCR) based on the amplification of specific classes of repeating elements, such as long dispersed nuclear element-1 (LINE-1) (P Tangkijvanich et al 2007 Clin Chim Acta 379:127-133). One possible explanation for this observation is that although hypomethylation is widespread in the tumor genome, it still exhibits some degree of heterogeneity across genomic regions.

In fact, the authors of the present invention observed that the average values of plasma methylation densities in reference subjects varied throughout the genome (Fig. 56). Each red dot in Fig. 56 reflects the average value of the methylation density of one 1 Mb region in 32 healthy subjects. The graph shows all analyzed 1 Mb regions throughout the genome. The number in each box represents the chromosome number. The authors of the present invention observed the variation in the average values of the density of methylation from site to site.

The diagnostic algorithm of a simple PCR-based analysis does not allow for such heterogeneity between areas. This heterogeneity extends the range of methylation densities by

- 38 042157 observed in healthy individuals. Therefore, a more significant reduction in methylation density may be required for a sample to be considered as showing hypomethylation. This will reduce the sensitivity of the test.

In contrast, using the massively parallel sequencing method, the genome is divided into 1 Mb regions (or regions of other sizes) and the methylation densities for these regions are measured individually. This method reduces the impact of variations in the initial methylation densities in different genomic regions, since each region is compared in the test sample and in controls. Thus, for the same site, the variability between 32 healthy controls was relatively small. For 32 healthy controls, 95% of the plots had a coefficient of variation (CV) of <1.8%. However, to further increase sensitivity in detecting cancer-associated hypomethylation, comparisons can be made across multiple genomic regions. Sensitivity will increase when testing multiple genomic regions, as this ensures that there is no effect of biological variability when no hypomethylation of the cancer sample for a particular region is observed when testing just one region.

A method involving comparing the methylation density of equivalent genomic regions in controls and test samples (e.g., testing each genomic region separately, and then possibly combining such results) and performing said comparison for multiple genomic regions results in a higher signal-to-noise ratio in detection hypomethylation associated with cancer. This method, including massively parallel sequencing, is presented for illustrative purposes. It can be expected that other methods that allow determining the methylation densities of several genomic regions and comparing the methylation densities of the corresponding regions in controls and test samples will provide a similar effect. For example, to achieve the desired effect, hybridization probes or inverted molecular probes can be designed that can target plasma DNA molecules originating from specific genomic regions, as well as determine the methylation level of the specified region.

In yet another embodiment, the sum of the Z-scores of all regions can be used to determine the presence of cancer or to monitor sequential changes in plasma DNA methylation levels. Due to the general hypomethylated nature of tumor DNA, the sum of the Z-scores will be lower in plasma obtained from an individual with cancer than in healthy controls. The sum of Z-scores for samples of preoperative and postoperative plasma from a patient with HCC was -49843.8 and -3132.13, respectively.

In other embodiments, other methods may be used to measure the level of plasma DNA methylation. For example, the proportion of methylated cytosine residues in total cytosine residues can be determined using mass spectrometry (ML Chen et al. 2013 Clin Chem; 59: 824-832) or massively parallel sequencing. However, since most cytosine residues are outside the context of CpG dinucleotides, the proportion of methylated cytosines in the total number of cytosine residues will be relatively small compared to the methylation levels determined in the context of CpG dinucleotides. The authors of the present invention determined the level of methylation of tissue and plasma samples obtained from a patient with HCC, as well as four plasma samples obtained from healthy controls. Methylation levels were measured in the context of CpG, any cytosines, in the contexts of CHG and CHH using whole genome massively parallel sequencing data. H refers to adenine, thymine, or cytosine residues.

Fig. 26D is Table 2660 showing methylation levels in a tumor tissue sample and preoperative plasma, intersecting with some of the plasma controls in the context of HH and CHG. Methylation levels in the tumor tissue sample and preoperative plasma were consistently lower compared to buffy coat, non-tumor liver tissue, postoperative plasma sample, and plasma samples of healthy controls, for both CpG and any cytosines. However, data based on methylated CpGs, i.e. methylation densities showed a wider dynamic range than data based on methylated cytosines.

In other embodiments, plasma DNA methylation status can be determined by anti-methylated cytosine antibody methods, such as methylated DNA immunoprecipitation (MeDIP). However, the accuracy of these methods is expected to be lower than the accuracy of sequencing-based methods due to variability in antibody binding. According to another implementation variant, the level of 5-hydroxymethylcytosine in plasma DNA can be determined. However, a decrease in the level of 5-hydroxymethylcytosine has been found to be an epigenetic feature of some cancers, such as melanoma (CG Lian, et al. 2012 Cell; 150: 1135-1146).

In addition to HCC, the authors of the present invention also investigated whether the application of this method for other types of cancers. The authors of the present invention examined plasma samples from 2 patients with adenocarcinoma of the lung (AL1 and AL2), 2 patients with nasopharyngeal

- 39 042157 carcinomas (NGK1 and NGK2), 2 patients with colon and rectal cancer (CRK1 and CRK2), 1 patient with metastatic neuroendocrine tumor (NEC1) and 1 patient with metastatic smooth muscle sarcoma (GMS1). Plasma DNA from these subjects was converted with bisulfite and sequenced using the Illumina HiSeq2000 platform for 50 bp. at one end. The four healthy controls mentioned above were used as a reference group for analysis in these 8 patients. Used 50 p. sequence reads at one end. The entire genome was divided into sections of 1 Mb in size. Mean and SD of methylation density were calculated for each site using reference group data. These results for 8 cancer patients were then expressed as Z-scores representing the SD number of the mean of the reference group. A positive value means that the methylation density in the test case is lower than the average value for the reference group, and vice versa. The number of sequencing reads and the achieved sequencing depth per sample are shown in Table 1. 2780 in FIG. 27I.

In FIG. 27A-H are Circos plots of methylation density in 8 cancer patients according to embodiments of the present invention. Each dot represents a result for a 1 MB chunk. Black dots indicate areas with Z-scores ranging from -3 to 3. Red dots indicate areas with Z-scores <-3. Green dots indicate areas with Z-scores >3. The spacing between two consecutive lines represents the Z-score difference of 20.

Significant hypomethylation has been observed in several regions of the genomes of patients with most types of cancer, including lung cancer, nasopharyngeal carcinoma, colon and rectal cancer, and metastatic neuroendocrine tumor. Interestingly, in addition to hypomethylation, significant hypermethylation of several regions across the genome was observed in the case of metastatic smooth muscle sarcoma. The smooth muscle sarcoma originates from the embryonic mesoderm, while the other types of cancers in the remaining 7 patients are of ectodermal embryonic origin. Thus, it is possible that the DNA methylation pattern of sarcoma may differ from that of carcinoma.

As can be seen from this case, the plasma DNA methylation pattern may also be suitable for differentiating different types of cancers, in this example, for differentiating between carcinoma and sarcoma. These data also suggest that this method can be used to detect aberrant hypermethylation associated with malignancy. In all of these 8 cases, only plasma samples were available and no tumor tissue analysis was performed. This showed that even in the absence of a prior methylation profile or methylation levels of tumor tissue, tumor-derived DNA could be easily detected in plasma using the methods described.

Fig. 27J is a table. 2790, which reflects the distribution of Z-scores of 1 Mb regions of the whole genome in the plasma of patients with various malignant neoplasms. For each case, the percentage of plots with a Z-score of <-3, -3-3, and >3 is shown. More than 5% of plots had a Z-score <-3 for all cases. Thus, when the present inventors use a cut-off point of 5% significantly hypomethylated areas to classify a sample as cancer-positive, all of these cases will be classified as cancer-positive. The results obtained by the authors show that hypomethylation is supposedly a common phenomenon for different types of cancers, and plasma methylome analysis is suitable for the detection of different types of cancers.

D. Method.

Fig. 28 is a flowchart of a method 2800 for analyzing a biological sample of an organism to establish a cancer level classification, in accordance with embodiments of the present invention. The specified biological sample includes DNA derived from normal cells, and may potentially include DNA from cells associated with cancer. At least a portion of the DNA in a biological sample may be free of cells.

At step 2810, a plurality of DNA molecules from the specified biological sample are analyzed. Analysis of a DNA molecule may include determining the localization of said DNA molecule within the genome of said organism and determining whether said DNA molecule is methylated at one or more sites. Said analysis can be performed by obtaining reads of sequences using methylation-sensitive sequencing, and, accordingly, the analysis can be performed solely on the basis of data previously obtained for said DNA. In other embodiments, the analysis may include actual sequencing or other active data acquisition steps.

At 2820, for each of the plurality of sites, the relative amount of DNA molecules that are methylated at that site is determined. In one embodiment, said sites are CpG sites, and may be only some CpG sites selected using one or more of the criteria mentioned herein. The number of DNA molecules that are methylated is equivalent to determining the number of unmethylated molecules if nor

- 40 042157 sequencing is performed using the total number of DNA molecules analyzed at a particular site, for example, the total number of sequencing reads. For example, an increase in the density of CpG methylation in a region is equivalent to a decrease in the density of unmethylated CpG in the same region.

At step 2830, the first level of methylation is calculated based on the respective amounts of DNA molecules methylated at the plurality of sites. The first level of methylation may correspond to the density of methylation, which is determined based on the number of DNA molecules corresponding to the specified set of sites. These sites may correspond to multiple loci or only one locus.

At 2840, the first methylation level is compared with the first cutoff. The first limit value may be a reference methylation level or be related to a reference methylation level (eg, be at a certain distance from the normal level). The reference level of methylation can be determined from samples of individuals without cancer, or from loci in the body that are not known to be associated with cancer in the specified organism. The first limit value can be set to a reference level of methylation, determined from a previous biological sample from a specified organism, obtained from a previously tested biological sample.

In one embodiment, the first cutoff value is a specified distance (eg, a specified number of standard deviations) from a reference methylation level established from a biological sample obtained from a healthy organism. The comparison can be performed by determining the difference between the first methylation level and the reference methylation level, and then comparing this difference with a threshold corresponding to the first limit value (for example, to determine if the methylation level is statistically different from the reference methylation level).

At 2850, a cancer level is classified based on said comparison. Examples of the level of cancer include the presence or absence of a cancerous disease or a premalignant condition in a subject, or an increased likelihood of developing a cancerous disease. In one embodiment, the first cut-off value may be determined from a previously obtained sample from the subject (eg, the reference methylation level may be determined from a previously obtained sample).

In some embodiments, the first methylation level may correspond to a number of regions where methylation levels exceed a threshold. For example, multiple regions of the genome of the specified organism can be identified. These areas can be identified using the criteria mentioned herein, for example, by a certain length or a certain number of sites. Within each of these regions, one or more sites (eg, CpG sites) can be identified. For each region, the methylation level of the region can be calculated. The first level of methylation corresponds to the first region. Each of the region's methylation levels is compared to a respective region cut-off value, which may be the same or vary from region to region. The limit value for the first region is called the first limit value. Appropriate region cutoffs can be represented by a specified amount (eg, half (0.5)) of the reference methylation level, so that only regions significantly different from a reference that can be determined from non-cancer subjects are counted.

A first number of regions whose methylation level exceeds the respective region limit may be determined and compared to a threshold value to determine a classification. In one embodiment, the threshold value is a percentage. Comparing said first number to a threshold may include dividing the first number of regions by a second number of regions (eg, all regions) prior to comparison with the threshold, eg, as part of a normalization process.

As described above, the fractional concentration of tumor DNA in a biological sample can be used to calculate the first limit value. A fractional concentration may simply be defined as being above a minimum value, while a sample with a fractional concentration below the minimum value may be flagged, for example, as unsuitable for analysis. The minimum value can be determined based on the expected difference in tumor methylation levels relative to the reference level of methylation. For example, if the difference is 0.5 (eg, used as a cut-off value), a specific tumor concentration is required that is high enough for said difference to occur.

The specific techniques of method 1300 can also be used in method 2800. According to method 1300, copy number variation can be determined for a tumor (eg, a first chromosomal region of a tumor can be tested for copy number variation relative to a second chromosomal region of the tumor). Accordingly, method 1300 may assume that a tumor exists. According to method 2800, a sample can be tested for any tumor in principle, regardless of any copy number characteristics. Some of the techniques of these two methods may be similar. However, cut-offs and methylation parameters (e.g. norms

- 41 042157 methylation levels) according to method 2800 detect statistical difference from the reference methylation level for non-tumor DNA, and not the difference from the reference methylation level for a mixture of cancer DNA and non-tumor DNA, with some regions possibly containing copy number variations. Accordingly, reference values for method 2800 can be determined from non-cancerous DNA samples, such as those obtained from organisms that do not have cancer, or from non-tumor tissue samples from the same patient (for example, plasma obtained earlier or obtained at the same time). time to samples known not to be affected by cancer, which can be determined from cellular DNA).

E. Prediction of the minimum fractional concentration of tumor DNA detected using plasma DNA methylation assay.

One way to measure the sensitivity of a cancer detection method using plasma DNA methylation is related to the minimum fractional concentration of tumor-derived DNA required to detect a change in plasma DNA methylation when compared to levels of controls. The sensitivity of the test also depends on the degree of DNA methylation difference between tumor tissue and baseline levels of plasma DNA methylation in healthy controls or blood cell DNA. Blood cells are the main source of DNA in the plasma of healthy individuals. The greater this difference, the easier it is to distinguish cancer patients from cancer-free individuals, which is reflected in lower detection thresholds of tumor-derived DNA in plasma and higher clinical sensitivity in detecting cancers in patients. In addition, variations in plasma DNA methylation in healthy subjects or subjects of different ages (G Hannum et al. 2013 Mol Cell; 49: 359-367) also affect the sensitivity of detecting methylation changes associated with the presence of cancer. Less variability in DNA methylation in the plasma of healthy subjects makes it easier to detect changes due to the presence of a small amount of DNA derived from cancer.

Fig. 29A is a graph 2900 depicting the distribution of methylation densities in reference subjects, assuming that the indicated distribution follows a normal distribution. This analysis is based on the presence of only one indicator of methylation density for each plasma sample, for example, the methylation density of all autosomes or a particular chromosome. It is shown how this affects the specificity of the analysis. In one embodiment, a cut-off point 3 SD below the mean DNA methylation density in reference subjects is used to determine if a test sample has significantly more hypomethylation than samples from reference subjects. Using this cut-off point, approximately 0.15% of cancer-free subjects are expected to have false positive results confirming the presence of cancer, corresponding to a specificity of 99.85%.

Fig. 29B is a graph 2950 depicting methylation density distributions in reference subjects and cancer patients. The limit value is 3 SD below the mean value of methylation densities in reference subjects. If the reported mean methylation densities of cancer patients are 2 SD below the cut-off value (i.e., 5 SD below the mean for reference subjects), 97.5% of cancer patients are expected to have methylation densities below the cut-off . In other words, the expected sensitivity is 97.5% if one methylation density is known for each subject, for example, if the total methylation density of the entire genome, all autosomes, or a particular chromosome is analyzed. The difference in mean methylation densities for the two populations is influenced by two factors, namely the degree of difference in methylation levels between cancerous and non-cancerous tissues, and the fractional concentration of tumor-derived DNA in the plasma sample. The higher the values of these two parameters, the greater the difference between the methylation densities for these two populations.

In addition, the lower the SD for the methylation density distributions for the two populations, the less the methylation density distributions for the two populations overlap.

In the present description, the authors use a hypothetical example to illustrate this principle. Let us assume that the methylation density of tumor tissue is approximately 0.45, and the plasma DNA methylation density of healthy subjects is approximately 0.7. These predicted values are similar to those obtained for our HCC patient, with a total autosomal methylation density of 42.9% and a mean autosomal methylation density for plasma samples from healthy controls of 71.6%. Assuming that the CV of whole genome plasma DNA methylation density measurement is 1%, the cut-off value would be 0.7x(100%-3x1%)=0.679. To achieve a sensitivity of 97.5%, the mean plasma DNA methylation density for cancer patients should be approximately 0.679-0.7x(2x1%)=0.665. Let f represent the fractional concentration

- 42 042157 tumor-derived DNA in a plasma sample. Then f can be calculated as (0.7-0.45)xf=0.7-0.665. Thus, f is approximately 14%. This calculation determines that the minimum fractional concentration that can be detected in plasma is 14% to provide a diagnostic sensitivity of 97.5% when total genome methylation density is used as the diagnostic parameter.

The present inventors then performed said analysis on data obtained from a patient with HCC. For illustrative purposes, in this case, only one measurement of the density of methylation was carried out, based on the indicator determined for all autosomes, for each sample. The mean methylation density was 71.6% in plasma samples from healthy subjects. The SD of the methylation densities for these four samples was 0.631%. Thus, the cut-off value for plasma methylation density should be 71.6%3x0.631%=69.7% to provide a Z-score <-3 and a specificity of 99.85%. To achieve a sensitivity of 97.5%, the mean plasma methylation density of cancer patients should be 2 SD below the cut-off point, i.e. be 68.4%. Since the methylation density of the tumor tissue was 42.9%, based on the formula P=BKGx(1-f)+TUMxf ₎ f should be at least 11.1%.

In another embodiment, the methylation densities of different genomic regions can be analyzed separately, for example, as shown in FIG. 25A or B. In other words, several measurements of the level of methylation were carried out for each sample. As shown below, significant hypomethylation can be detected at a significantly lower fractional concentration of tumor DNA in plasma and, accordingly, the diagnostic performance of plasma DNA methylation analysis for the detection of cancers will be higher. The number of genomic regions showing a significant deviation in methylation densities from the reference population can be counted. This number of genomic regions can then be compared to a cut-off value to determine whether there is significant total plasma DNA hypomethylation in the population for the genomic regions under study, eg, 1 Mb regions across the genome. The limit value can be established by analyzing a group of reference subjects without cancer, or derived mathematically, for example, in accordance with the normal distribution function.

Fig. 30 is a graph 3000 showing the distribution of plasma DNA methylation densities in healthy subjects and cancer patients. The methylation density of each 1 Mb region is compared with the corresponding values for the reference group. The percentage of sites showing significant hypomethylation (3 SD below the mean for the reference group) was determined. A cut-off point of 10% of significantly hypomethylated areas was used to determine the presence of tumor-derived DNA in a plasma sample. Other cutoffs such as 5%, 15%, 20%, 25%, 30%, 35%, 40%, 45%, 50%, 60%, 70%, 80%, or 90% may also be used, according to the required sensitivity and specificity of the test.

For example, to classify a sample as containing tumor-derived DNA, the present inventors may use 10% of 1 Mb regions showing significant hypomethylation (Z-score <-3) as a cutoff point. If more than 10% of these areas are significantly more hypomethylated than the reference group, the sample is classified as positive in a cancer test. For each 1 Mb region, a cut-off point 3 SD below the mean methylation density of the reference group is used to define the sample as significantly more hypomethylated. For each 1 Mb region, if the mean plasma DNA methylation density of cancer patients is 1.72 SD lower than the mean plasma DNA methylation density of reference subjects, there is a 10% chance that the methylation density of any particular region in a patient with cancer will be lower than the cut-off point (i.e. Z-score <-3) and give a positive result. In such a case, when considering all 1 Mb regions for the whole genome, approximately 10% of the regions are expected to show positive results at significantly lower methylation densities (i.e., Z-scores <-3). Assuming that the total plasma DNA methylation density of healthy subjects is approximately 0.7, and the coefficient of variation (CV) of the plasma DNA methylation density measurement for each 1 Mb region is 1%, the average plasma DNA methylation density of cancer patients should be 0.7x(100%1.72x1%) = 0.68796. Let f be the fractional plasma concentration of tumor-derived DNA that provides the indicated mean plasma DNA methylation density. Assuming a tumor tissue methylation density of 0.45, the f value can be calculated using the following equation:

_ (Mp^ -M _lumor )xf = Mp^ -Mp^ where ^Pr 'f represents the mean plasma DNA methylation density in the reference

- 43,042,157 individuals;

Η"™,· represents the density of methylation of the tumor tissue in a patient with a cancer patient;

and Mp^'r represents the mean plasma DNA methylation density in cancer patients.

Based on the above equation (0.7-0.45)xf=0.7-0.68796. Accordingly, the minimum fractional concentration that can be detected using this method is defined as 4.8%. Sensitivity can be further improved by reducing the cutoff percentage of significantly more hypomethylated regions, for example from 10 to 5%.

As shown in the above example, the sensitivity of this method is determined by the degree of difference in the level of methylation of cancerous and non-cancerous tissues, such as blood cells. In one embodiment, only chromosomal regions are selected that show a significant difference in DNA methylation densities in the plasma of non-cancer subjects and DNA methylation densities in tumor tissue. In one embodiment, only regions are selected whose methylation density difference is >0.5. In other embodiments, a difference of 0.4, 0.6, 0.7, 0.8, or 0.9 may be used to select suitable regions. In other embodiments, the physical size of the genomic regions is not fixed. Instead, genomic regions are determined, for example, based on a fixed sequencing depth or a fixed number of CpG sites. Methylation levels in several indicated genomic regions are assessed for each sample.

Fig. 31 is a graphical representation 3100 showing the distribution of differences in mean DNA methylation densities between plasma of healthy subjects and tumor tissue of an HCC patient. A positive value means that the methylation density is higher for DNA in the plasma of healthy subjects, and a negative value means that the methylation density is higher in tumor tissue.

In one embodiment, sites with a maximum difference in methylation density between cancerous and non-cancerous tissues, such as those for which the difference is >0.5, regardless of whether the tumor is hypomethylated or hypermethylated at said sites, can be selected. The limit of detection of fractional concentration of tumor-derived DNA in plasma can be reduced by focusing on these areas, due to the greater differences in the distributions of plasma DNA methylation levels between cancer patients and non-cancer patients, with the same fractional concentration of tumor-derived DNA in plasma. For example, by using only regions with a difference >0.5, and assuming a cut-off point corresponding to 10% significantly more hypomethylated regions to determine the presence of cancer in a test individual, the minimum fractional concentration (f) of detectable tumor-derived DNA can be calculated using following equation:

{Mpn _f - M^) xf = Mpn _f - Mp^

MP

Where ^r represents the mean plasma DNA methylation density in reference individuals;

tumor represents the methylation density of tumor tissue in a cancer patient;

~Mp represents the mean plasma DNA methylation density in cancer patients.

In this case, the difference in methylation density in the plasma of reference subjects and in tumor tissues is at least 0.5. Accordingly, we have 0.5xf= 0.7-0.68796 and f=2.4%. Thus, by focusing on areas with a greater difference in methylation density between cancerous and non-cancerous tissues, the lower limit for the fractional concentration of tumor-derived DNA can be reduced from 4.8 to 2.4%. Information as to which regions show large differences in methylation between cancerous and non-cancerous tissues, such as blood cells, can be obtained using tumor tissues from the same organ or tissues of the same histological type obtained from other individuals.

In another embodiment, based on the plasma DNA methylation density for all regions, a parameter can be obtained that takes into account the difference in methylation densities between cancerous and non-cancerous tissues. Regions with a greater difference may be given more weight. In one embodiment, the difference in methylation density between cancerous and non-cancerous tissue for each site can be directly used as the site-specific weight in calculating the final parameter.

In another embodiment, different types of cancers may have different

- 44 042157 methylation patterns in tumor tissue. A cancer-specific weight profile can be obtained based on the degree of methylation in a particular type of cancer.

According to another implementation variant, the dependence of the density of methylation between sites in subjects who have or do not have cancer can be determined. In FIG. 8, it can be observed that in a small number of areas, tumor tissues are more methylated than plasma DNA in reference subjects. Accordingly, areas with extreme difference values, for example, differences >0.5 and <0, can be selected. The ratio of methylation density of these sites can then be used to determine the presence of cancer in the test individual. In other embodiments, the difference and quotient of methylation density of different sites can be used as parameters to determine the relationship between sites.

The authors also evaluated the sensitivity of a method for detecting or assessing a tumor using methylation densities of several genomic regions, which is illustrated by data obtained from a patient with HCC. First, the present inventors combined reads from preoperative plasma with reads from healthy control plasma samples to mimic plasma samples containing tumor DNA fractions ranging from 20% to 0.5%. The authors then calculated the percentage of 1 Mb regions (out of 2734 regions of the entire genome) with methylation densities equivalent to Z-scores <-3. At a fractional concentration of tumor DNA in plasma of 20%, significant hypomethylation was observed for 80.0% of the sites. The corresponding data for plasma tumor DNA fractions of 10%, 5%, 2%, 1%, and 0.5% were 67.6%, 49.7%, 18.9%, 3.8%, and 0.77% of sites showing hypomethylation, respectively. Since the theoretical limit for the number of regions showing Z-scores <-3 in controls is 0.15%, our data demonstrate that even at a fractional concentration of tumor DNA of only 0.5%, there were more regions (0.77 %) beyond the theoretical cutoff.

Fig. 32A is a table. 3200 showing the effect of reducing the depth of sequencing when the plasma sample contained 5% or 2% tumor DNA. A high content of regions (>0.15%) showing significant hypomethylation can be detected even at an average sequencing depth of only 0.022x of the haploid genome.

Fig. 32B is a graphical representation 3250 plotting the methylation densities of repeat elements and regions without repeats in the plasma of four healthy controls, buffy coat, normal liver tissue, tumor tissue, preoperative plasma, and postoperative plasma from an HCC patient. It can be seen that repeat elements are more methylated (higher methylation density) than regions without repeats in both cancerous and non-cancerous tissues. However, the difference in methylation between repeat elements and regions without repeats was more significant in non-cancerous tissues and plasma DNA of healthy subjects, compared with tumor tissues.

As a result, plasma DNA of a cancer patient showed a greater decrease in methylation density for repeat elements than for regions without repeats. The difference in plasma DNA methylation density between the mean of the four healthy controls and the HCC patient was 0.163 and 0.088 for repetitive elements and non-repetitive regions, respectively. Data for preoperative and postoperative plasma samples also showed that the dynamic range of methylation density was greater for repetitive regions compared to non-repeated regions. In one embodiment, the methylation density of repetitive DNA elements in plasma can be used to determine if a patient has cancer or to monitor disease progression.

As discussed above, variations in plasma methylation densities in reference subjects also affect the accuracy of differentiating cancer patients from cancer-free individuals. The denser the distribution of methylation densities (ie, the smaller the standard deviation), the more accurate will be the differentiation between cancer patients and non-cancer subjects. In another embodiment, the coefficient of variation (CV) for methylation densities of 1 Mb regions can be used as a criterion for selecting regions with low variability in plasma DNA methylation densities in a reference group. For example, only sites with CV<1% are selected. Other values such as 0.5%, 0.75%, 1.25% and 1.5% can also be used as criteria for selecting regions of low methylation density variability. According to another implementation variant, the selection criteria may include both the CV for the specified site, and the difference in methylation density between cancerous and non-cancerous tissues.

The methylation density can also be used to determine the fractional concentration of tumor-derived DNA in a plasma sample if the methylation density of the tumor tissue is known. This information can be obtained by analyzing a patient's tumor or examining the tumors of several patients who have the same type of cancer. As discussed above, plasma methylation density (P) can be expressed using the following equation:

- 45 042157

where BKG is the background methylation density in blood cells and other organs, TUM is the methylation density in tumor tissue, and f is the fractional concentration of tumor-derived DNA in the plasma sample. This equation can be reformulated as:

BKG-P? ~ BKG-TUM

BKG values can be determined by analyzing a patient's plasma sample at a time point when cancer is absent, or by examining a reference group of individuals without cancer. Thus, after measuring the density of methylation in plasma, the value of f can be determined.

F. Combination with other methods.

The methylation assay methods described herein can be used in combination with other methods that rely on genetic alterations in plasma of tumor-derived DNA. Examples of such methods include analysis for cancer-associated chromosomal aberrations (KCA Chan et al. 2013 Clin Chem; 59:211-224; RJ Leary et al. 2012 Sci Transl Med; 4:162ra154) and cancer-associated single nucleotide variations in plasma (KCA Chan et al. 2013 Clin Chem; 59:211-224). The use of methylation analysis has some advantages over these genetic methods.

As shown in FIG. 21A, tumor DNA hypomethylation is a global phenomenon spanning regions distributed almost throughout the genome. Thus, DNA fragments from all chromosomal regions would be informative in terms of potentially contributing to the presence of tumor-derived hypomethylated DNA in the patient's plasma/serum. At the same time, chromosomal aberrations (amplification or division of a chromosomal region) are present only in some chromosomal regions, and DNA fragments from regions without chromosomal aberrations in tumor tissue will not be informative for analysis (KCA Chan et al. 2013 Clin Chem; 59: 211- 224). Similarly, only a few thousand single nucleotide changes are observed in each cancer genome (KCA Chan et al. 2013 Clin Chem; 59: 211-224). DNA fragments that do not overlap with the indicated single nucleotide changes will not be informative for determining the presence of DNA in plasma originating from a tumor. Thus, the proposed method for analyzing methylation is potentially a more cost-effective approach than these genetic methods for detecting cancer-related changes in blood flow.

In one embodiment, the cost-effectiveness of plasma DNA methylation analysis can be further increased by enriching DNA fragments from the most informative regions, eg, regions with the highest methylation differences between cancerous and non-cancerous tissues. Examples of methods for enriching these regions include the use of hybridization probes (eg, Nimblegen SeqCap systems and Agilent SureSelect Target Enrichment systems), PCR amplification, and solid phase hybridization.

G. Tissue-specific analysis/donors.

Tumor-derived cells invade and metastasize to adjacent or distant organs. Affected tissues or metastases contribute to plasma DNA through cell death. By analyzing the DNA methylation profile in the plasma of cancer patients and detecting the presence of tissue-specific methylation signatures, tissue types that are involved in the pathological process can be detected. This method provides a non-invasive anatomical scan of the tissues involved in the malignant process, which helps to identify the organs involved in the primary lesions and metastatic sites. Monitoring the relative plasma concentrations of the methylation signatures of the organs involved will also allow assessment of the tumor burden in said organs and determine whether cancer-related damage in said organ is progressing, or improving or being cured. For example, suppose that gene X is specifically methylated in the liver. In this case, it can be expected that when a cancer (eg, colon and rectal cancer) metastasizes to the liver, the concentration of methylated gene X sequences in plasma will increase. There may also be other sequence or groups of sequences with similar methylation characteristics to gene X. In such a case, the results for such sequences may be combined. Similar considerations apply to other tissues such as the brain, bones, lungs and kidneys, and so on.

On the other hand, tissue-specific methylation signatures are known to be observed in DNA from different organs (BW Futscher et al. 2002 Nat Genet; 31:175-179; SSC Chim et al. 2008 Clin Chem; 54: 500-511). Accordingly, plasma methylation profiling can be used to determine the contribution made by tissues of various organs to plasma. The establishment of such a contribution can be used to assess organ damage, since it is believed that plasma DNA is released when cells die. For example, liver disease such as hepatitis (eg caused by viruses, autoimmune processes, etc.) or hepatotoxicity (eg overdose of drugs (eg paracetamol) or toxins (eg alcohol)) caused by drugs,

- 46 042157 is associated with damage to liver cells and is expected to be associated with an increase in plasma levels of liver-derived DNA. For example, suppose that gene X is specifically methylated in the liver. In this case, it can be expected that in liver pathology, the concentration of methylated gene X sequences in plasma will increase. Conversely, suppose that the Y gene is specifically hypomethylated in the liver. In this case, it can be expected that the concentration of methylated Y gene sequences in plasma will decrease in liver pathology. In yet another embodiment, the X or Y gene can be replaced with any genomic sequence, which is not necessarily a gene that exhibits methylation differences in different body tissues.

The techniques described herein can also be used to evaluate donor-derived DNA in the plasma of organ transplant recipients (YMD Lo et al. 1998 Lancet; 351:13291330). Donor-recipient polymorphism differences have been used to differentiate donor-derived and recipient-derived DNA in plasma (YW Zheng et al. 2012 Clin Chem; 58: 549-558). The authors of the present invention suggested that tissue-specific methylation signatures of a transplanted organ can also be used as a method for detecting donor DNA in recipient plasma.

By monitoring the concentration of donor DNA, the status of the transplanted organ can be assessed in a non-invasive way. For example, transplant rejection is associated with a higher rate of cell death and, accordingly, the concentration of donor DNA in the recipient's plasma (or serum), as reflected by the methylation signature of the transplanted organ, will be increased relative to the concentration at the time the patient's condition is stable, or compared to other stable transplant recipients or healthy controls who have not received transplantation. Similar to what has been described for cancers, donor-derived DNA can be identified in the plasma of transplant recipients by detecting all or some of the characteristic features, including polymorphism differences, of smaller DNA for transplanted solid organs (YW Zheng et al. 2012 Clin Chem; 58: 549-558) and a tissue-specific methylation profile.

H. Normalization of methylation by size.

According to the description above and that of Lun et al (FMF Lun et al. Clin. Chem. 2013; CIO:10.1373/clinchem.2013.212274), methylation density (eg, plasma DNA) correlates with the size of DNA fragments. The distribution of methylation densities for shorter plasma DNA fragments was significantly lower than for longer fragments. The present inventors contemplate that in some non-malignant conditions (e.g., systemic lupus erythematosus (SLE)) with abnormal plasma DNA fragmentation patterns, there may be apparent hypomethylation of plasma DNA due to the presence of more short plasma DNA fragments that are less methylated. In other words, plasma DNA size distribution may be a confounding factor for plasma DNA methylation density.

In FIG. 34A shows the size distribution of plasma DNA in a patient with SLE SLE04. The size distributions for nine healthy controls are shown as dotted gray lines and the distributions for SLE04 are shown as a solid black line. Short plasma DNA fragments were present in higher numbers in SLE04 compared to nine healthy controls, because shorter DNA fragments are generally less methylated, this size distribution pattern may confound the plasma DNA methylation assay and result in more hypomethylation.

In some embodiments, the measured methylation level can be normalized to reduce the confounding effect of size distribution on plasma DNA methylation analysis. For example, the size of DNA molecules at multiple sites can be measured. In various embodiments, said measurement may provide a specific size (eg, length) of a DNA molecule, or simply determine that the size falls within a specific range, which may also match the size. The normalized level of methylation can then be compared with the limit value. There are several ways to perform normalization to reduce the confounding effect of size distribution on plasma DNA methylation analysis.

According to one implementation variant, fractionation according to the size of DNA (eg, plasma DNA) can be carried out. Size fractionation can ensure that DNA fragments of a similar size are used to determine the level of methylation according to the cut-off value. Within size fractionation, DNA fragments of a first size (eg, a first length range) can be selected, with the first limit corresponding to the first size. Normalization can be done by calculating the level of methylation using only selected DNA fragments.

Size fractionation can be achieved in various ways, for example, either by physically separating DNA molecules of different sizes (eg, by electrophoresis, or by microfluidics or centrifugation-based technologies), or by in silico analysis. For in silico analysis according to one implementation variant, a

- 47 042157 massively parallel sequencing of paired ends of plasma DNA molecules. The size of the sequenced molecules can then be determined by comparing the location of each of the two ends of the plasma DNA molecule with a reference human genome. Subsequent analysis can then be performed by selecting sequenced DNA molecules that meet one or more size selection criteria (eg, criteria that determine size within a specified range). Accordingly, in one embodiment, methylation density can be analyzed for fragments of a similar size (eg, within the indicated range). A limit value (eg, at step 2840 of method 2800) may be determined based on fragments from the same size range. For example, methylation levels can be determined from samples known to be cancerous or non-cancer, and cutoffs can be determined from said methylation levels.

In another embodiment, a functional relationship between methylation density and circulating DNA size can be determined. Said functional dependency can be defined by a data point or function coefficients. This functional dependency may provide scaling values corresponding to the respective sizes (eg, methylation at shorter sizes may increase accordingly). According to various implementations, the scaling value may be from 0 to 1, or greater than 1.

You can normalize to the average size. For example, an average size corresponding to the DNA molecules used to calculate the first methylation level can be calculated, and said first methylation level can be multiplied by the appropriate scaling value (ie corresponding to the average size). In another example, the methylation density of each DNA molecule can be normalized according to the size of the DNA molecule and the relationship between size and DNA methylation.

According to another implementation variant, the normalization can be performed per molecule. For example, the appropriate size of the DNA molecule at a particular site can be obtained (eg, as described above), and the scaling value corresponding to the corresponding size can be identified from a functional relationship. In the non-normalized calculation, each molecule is counted in the same way when determining the methylation index at the specified site. In a normalized calculation, the contribution of a molecule to the methylation index can be weighted by a scaling factor that corresponds to the size of the molecule.

In FIG. 34B and C are methylation analyzes for plasma DNA from SLE patient SLE04 (FIG. 34B) and HCC patient TBR36 (FIG. 34C). Outer circles represent _Zmeth results for plasma DNA without in silico size fractionation. Inner circles represent Z _meth results for 130 bp plasma DNA. or more. In a SLE patient, SLE04, 84% of the sites were hypomethylated without size fractionation in silico. The percentage of sites showing hypomethylation decreased to 15% when only 130 bp fragments were analyzed. or more. In a patient with HCC TBR36, 98.5% and 98.6% of plasma DNA regions were hypomethylated with and without size fractionation in silico, respectively. These results suggest that in silico size fractionation is effective in reducing false positive hypomethylation assays associated with increased plasma DNA fragmentation, such as in patients with SLE or other inflammatory conditions.

In one embodiment, the results of assays with and without size fractionation can be compared to determine if there is any confounding effect of size on the results of the methylation assay. Accordingly, in addition to or instead of normalization, a size-specific methylation rate calculation can be used to determine whether there is a possibility of a false positive result if the percentage of plots above the cut-off value differs between size fractionation and non-size fractionation, or only the specific methylation level differs. For example, the presence of a significant difference between these results for samples with and without size fractionation can be used to determine the likelihood of a false positive result due to an abnormal fragmentation pattern. A threshold for determining whether a difference is significant can be established by analyzing a cohort of cancer patients and a cohort of non-cancer controls.

I. Analysis of genome-wide hypermethylation of CpG islands in plasma.

In addition to general hypomethylation, hypermethylation of CpG islands is also frequently observed in cancers (SB Baylin et al. 2011 Nat Rev Cancer; 11: 726-734; PA Jones et al. 2007, Cell; 128: 683-692; M Esteller et al 2007 Nat Rev Genet 2007; 8: 286-298; M Ehrlich et al. 2002 Oncogene 2002; 21: 5400-5413). In this section, the present inventors describe the use of a genome-wide analysis of CpG island hypermethylation for the detection and monitoring of cancers.

Fig. 35 is a flowchart of a method 3500 for determining cancer level classification based on CpG island hypermethylation, in accordance with embodiments of the present invention.

- 48 042157 shade. The plurality of sites according to method 2800 may include CpG sites, wherein said CpG sites are organized into a plurality of CpG islands, where each CpG island includes one or more CpG sites. The methylation levels for each CpG island can be used to determine the level of cancer classification.

At 3510, the CpG islands to be analyzed are identified. In this assay, as an example, the present inventors first identified a set of CpG islands for analysis characterized by relatively low methylation densities in the plasma of healthy reference subjects. In one aspect, the variation in methylation densities in a reference group may be relatively small, making it easier to detect cancer-associated hypermethylation. In one embodiment, the mean methylation density of the CpG islands is less than the first percent in the reference group, and the coefficient of variation for the methylation density in the reference group is less than the second percent.

For example, for purposes of illustration, the following criteria are used to identify suitable CpG islands.

i) Mean methylation density for the CpG island in the reference group (eg healthy subjects) <5%.

ii) Coefficient of variation for analysis of plasma methylation density for a reference group (eg healthy subjects) <30%.

The specified parameters can be adjusted for a particular application. In the data set obtained by the authors, 454 CpG islands in the genome met the specified criteria.

In step 3520, the methylation density is calculated for each CpG island. Methylation densities can be determined as described herein.

At step 3530, it is determined whether each of the CpG islands is hypermethylated. For example, to analyze for hypermethylation of a CpG island in a test case, the methylation density of each CpG island was compared with the corresponding data of the reference group. The density of methylation (an example of the level of methylation) can be compared with one or more limit values to determine if a particular islet is hypermethylated.

According to one implementation variant, the first limit value may correspond to the average value of the methylation densities for the reference group, plus the specified percentage. Another limit value may correspond to the mean value of methylation densities for the reference group plus the indicated number of standard deviations. According to one implementation variant, the Z-score (Z _meth ) was calculated and compared with the limit values. For example, a CpG island in a test subject (eg, a subject being screened for cancer) was considered significantly hypermethylated if it met the following criteria:

i) its methylation density was higher than the reference group mean by 2%, and ii) ^Z meth>3.

The specified parameters can also be adjusted for a particular application.

At 3540, the methylation densities (eg, as Z-scores) of the hypermethylated CpG islands are used to determine a cumulative score. For example, once all significantly hypermethylated CpG islands have been identified, a score including the sum of the Z scores or functions of the Z scores of all hypermethylated CpG islands can be calculated. An example of a metric is a cumulative probability (CP) metric, as described in another section. The sum probability measure uses Zmeth to determine the probability of such an observation due to a random effect according to a probability distribution (for example, a t student probability distribution with 3 degrees of freedom).

At 3550, the cumulative score is compared to a cumulative threshold to establish a classification of cancer level. For example, if the total hypermethylation in the identified CpG islands is high enough, the organism can be identified as having cancer. In one embodiment, the cumulative threshold corresponds to the maximum cumulative score of the reference group.

IX. Methylation and CNA (copy number aberrations)

As mentioned above, the methylation assay methods described herein can be used in combination with other methods that rely on genetic alterations in plasma of tumor-derived DNA. Examples of such methods include analysis for cancer-associated chromosomal aberrations (KCA Chan et al. 2013 Clin Chem; 59: 211-224; RJ Leary et al. 2012 Sci Transl Med; 4: 162ra154). Aspects of copy number aberrations (CNA) are described in US Patent Application No. 13/308473.

A. C.N.A.

Copy number aberrations can be detected by counting the DNA fragments that align with a particular part of the genome, normalizing the resulting count, and comparing the counted value to a limit value. According to various implementation options, normalization can be carried out by counting DNA fragments aligned to a different haplotype of the same part of the geno

- 49 042157 ma (method of analysis of relative doses of haplotype (RHDO)) or by counting DNA fragments aligned to another part of the genome.

The RHDO method is based on the use of heterozygous loci. The embodiments described in this section can also be applied to homozygous loci by comparing two regions rather than two haplotypes of the same region, and thus are not haplotype specific. In a related method of relative doses of chromosomal regions, the number of fragments of one chromosomal region (for example, determined by counting the reads of sequences aligned to the specified region) is compared with the expected value (which may refer to a reference chromosomal region or the same region in another, known to be healthy sample) . Thus, a fragment of a chromosomal region will be counted regardless of which haplotype the sequenced label comes from. Accordingly, sequence reads that do not contain heterozygous loci may still be used. To perform a comparison, according to an implementation variant, the number of labels counted may be normalized before said comparison. Each region is defined by at least two loci (which are separate from each other), and fragments at these loci can be used to obtain an overall value for the specified region.

A normalized value for sequencing reads (tags) for a particular region can be calculated by dividing the number of sequencing reads aligned with the specified region by the total number of sequencing reads aligned with the entire genome. The specified number of normalization marks count allows you to compare the results from one sample with the results from another sample. For example, the normalized value may be the proportion (eg, percentage or fraction) of sequencing reads believed to be from a particular region, as indicated above. In other embodiments, other normalization methods may be used. For example, normalization is possible by dividing the count for one region by the count for the reference region (in the case described above, the reference region is simply the whole genome). This counted number of normalization marks can then be compared to a threshold value that can be determined from one or more cancer-free reference samples.

The counted number of normalization marks of the test case is then compared to the counted number of normalization marks of one or more reference subjects, eg, cancer-free subjects. According to one implementation variant of the specified comparison is carried out by calculating the Z-score of the specified case for a particular chromosomal region. The Z-score can be calculated using the following equation: Z-score = (calculated number of normalization marks of the specified case - mean) / SD, where the mean is the average of the counted number of normalization marks aligned to a particular chromosomal region for the reference samples; and SD is the standard deviation of the counted number of normalization marks aligned to a particular area for the reference samples. Accordingly, the Z-score is the number of standard deviations by which the counted number of normalization marks for a chromosomal region for a test case differs from the mean of the counted number of normalization marks for the same chromosome region in one or more reference subjects.

In a situation where the test organism has a cancer, the chromosomal regions amplified in the tumor tissues will be over-represented in plasma DNA. This will result in a positive Z-score. On the other hand, chromosomal regions that are deleted in tumor tissues will be underrepresented in plasma DNA. This will result in a negative Z-score. The value of the Z-score is determined by several factors.

One factor is the fractional concentration of tumor-derived DNA in a biological sample (eg, plasma). The higher the fractional concentration of tumor-derived DNA in the sample (eg, plasma), the greater will be the difference between the counted number of normalization marks for the test case and the reference cases. Accordingly, a higher Z-score value will be observed.

Another factor is represented by the variation in the counted number of normalization marks in one or more reference cases. Given the same amount of chromosomal region in a biological sample (eg, plasma) for a test case, less variation (i.e., less standard deviation) in the counted number of normalization marks in the reference group will result in a higher Z-score. Similarly, for the same degree of underrepresentation of a chromosomal region in a biological sample (eg, plasma) for a test case, a smaller standard deviation of the counted number of normalization marks in the reference group will result in a more negative Z-score value.

Another factor is the magnitude of chromosomal aberrations in tumor tissues. The magnitude of chromosomal aberrations refers to changes in the number of copies for a particular chromosomal region (addition or loss). The more significant changes in the number of copies in tumor tissues, the higher the degree of overrepresentation or underrepresentation of a particular chromosomal region in

- 50 042157

Plasma DNA.

For example, the loss of both copies of a chromosome will result in a greater under-representation of the chromosomal region in plasma DNA than the loss of one of the two copies of a chromosome, and thus result in a more negative Z-score. As a rule, in cancers there are several chromosomal aberrations. These chromosomal aberrations in each cancer may further vary in nature (i.e., amplification or division), extent (one or more copy additions or losses), and extent (size of the aberration relative to chromosome length).

The measurement accuracy of the calculated number of normalization marks is affected by the number of analyzed molecules. The authors of the present invention suggest that analysis of 15,000, 60,000 and 240,000 molecules is necessary to detect chromosome aberrations with one change in the number of copies (addition or loss) if the fractional concentration is approximately 12.5%, 6.3% and 3.2 %, respectively. For more information on cancer detection labels for different chromosomal regions, see US Patent Publication No. 2009/0029377 by Lo et al. genome sequencing), the entire contents of which are incorporated herein by reference in all respects.

Embodiments may also use size analysis instead of the mark count method. Dimensional analysis can also be used instead of counting normalization marks. When analyzing dimensions, various parameters can be used, mentioned in this document and in US patent application No. 12/940992. For example, Q or F values can be used, see above. Such dimensional values do not need to be normalized by counts for other areas, since the indicated values do not change depending on the number of reads. Methods using haplotype-specific methods, such as the RHDO method described above and also described in more detail in US patent application No. 13/308473, can also be used for non-specific methods. For example, techniques involving sequencing depth and region refinement can be used. In some embodiments, the standard GC offset for a particular region may be considered if two regions are compared. Since the same area is used in the RHDO method, no such correction is needed.

Although some cancers may in general be accompanied by aberrations in specific chromosomal regions, such cancers are not always accompanied by aberrations exclusively in such regions. For example, aberrations in additional chromosomal regions may be observed and the location of such additional regions may not be known. In addition, when screening patients to detect early stages of cancer, it may be necessary to identify a wide range of cancers in which genome-wide aberrations can be observed. In these situations, embodiments may analyze multiple regions in a systematic manner to determine in which regions aberrations are observed. The number of aberrations and their location (for example, information about whether they are consecutive) can be used, for example, to confirm aberrations, determine the stage of cancer, diagnose cancer (for example, if the specified number is above a threshold value), and obtain a prediction based on the specified the number and localization of the various regions in which aberration is observed.

Accordingly, embodiments make it possible to identify the presence of a cancerous disease in an organism based on the number of regions in which an aberration is observed. Accordingly, a plurality of regions (eg, 3000) can be tested to identify the number of regions in which aberration occurs. These regions may cover the entire genome or only parts of the genome, such as a region without repeats.

Fig. 36 is a flow diagram of a method 3600 for analyzing a biological sample of an organism using multiple chromosomal regions, in accordance with embodiments of the present invention. Said biological sample includes nucleic acid molecules (also referred to as fragments).

At 3610, a plurality of regions (eg, non-overlapping regions) of the specified organism's genome are identified. Each chromosomal region includes many loci. The size of the area can be 1 MB, or be represented by some other equal size. In a situation where the size of the region is 1 Mb, the entire genome may include approximately 3000 regions, each of which has a predetermined size and location. Such predetermined regions may vary to suit the length of a particular chromosome, or the specified number of regions to be used, and any other criteria referred to herein. If the regions have different lengths, such lengths can be used to normalize the results, for example, as described herein. These areas may be specifically selected based on certain criteria for a particular organism and/or based on information about the cancer being tested. Areas can also be selected

- 51 042157 optional.

At step 3620, the location of the nucleic acid molecule in the reference genome of the indicated organism is identified for each of the plurality of nucleic acid molecules. Localization can be determined by any of the methods mentioned herein, for example, by sequencing fragments to obtain sequencing marks and aligning said sequenced marks to a reference genome. For haplotype-specific methods, the specific haplotype of the molecule can also be determined.

Steps 3630-3650 are performed for each of the chromosomal regions. At 3630, an appropriate group of nucleic acid molecules is identified as originating from a chromosomal region based on the identified locations. Said respective group may include at least one nucleic acid molecule located at each of the plurality of loci in the chromosomal region. In one embodiment, said group may be represented by fragments that align to a particular haplotype of a chromosomal region, such as in the RHDO method described above. According to another implementation variant of the specified group can be represented by any fragment that is aligned to the chromosomal region.

At step 3640, the computer system calculates the corresponding value for the corresponding group of nucleic acid molecules. Said respective value determines the properties of the nucleic acid molecules of the respective group. Said corresponding value may be any of the values referred to herein. For example, the specified value may be the number of fragments in a group or a statistic of the distribution of fragment sizes in a group. The corresponding value may also be a normalized value, such as the counted number of marks of the specified area divided by the total number of counted marks for the specified sample, or the counted number of marks for the reference area. The corresponding value may also be defined relative to another value as a difference or ratio (eg, in the RHDO method), thus defining a distinguishing feature for the region.

At 3650, the corresponding value is compared with a reference value to determine whether the first chromosomal region is classified as showing a deletion or amplification. Said reference value may be 109 any of the threshold or reference values described herein. For example, the reference value may be a threshold value determined for normal samples. In the RHDO method, the corresponding value may be the difference or ratio of the number of labels counted for the two haplotypes, and the reference value may be a threshold for determining if there is a statistically significant deviation. In another example, the reference value may be a score count or size value for another haplotype or region, and the comparison may include obtaining a difference or ratio (or function thereof) and then determining if said difference or ratio exceeds a threshold value.

The reference value may vary depending on the results for other areas. For example, if adjacent areas also exhibit deviation (albeit small relative to one threshold, such as a Z-score of 3), a lower threshold may be applied. For example, if the first threshold is exceeded for three consecutive areas, cancer may be more likely. Accordingly, said first threshold may be lower than another threshold needed to identify cancer in non-consecutive areas. The presence of three regions (or more than three) showing even a slight deviation may have a sufficiently low probability of a random effect to maintain sensitivity and specificity.

At 3660, the number of genomic regions classified as showing a deletion or amplification is determined. The chromosomal regions that can be counted may be limited. For example, only regions that are consecutive with respect to at least one other region may be counted (or a certain size of consecutive regions may be required, such as 4 or more regions). In embodiments with unequal regions, said number may also be determined in terms of respective lengths (eg, said number may be the total length of the aberrant regions).

At 3670, the specified amount is compared to a threshold amount to determine the classification of the sample. For example, said classification may determine the presence of a cancer in the body, the stage of the cancer, and the prognosis for the cancer. In one embodiment, all aberrant regions are counted and a single threshold value is used regardless of the location of said regions. According to another implementation variant, the threshold value may vary based on the location and size of the areas to be counted. For example, the number of regions on a particular chromosome or chromosome arm may be compared to a threshold for that particular chromosome (or arm). Multiple thresholds may be used. For example, the number of aberrant regions on a particular chromosome (or arm) must be greater than the first threshold, and the total number of aberrant regions in the genome must be greater than the second threshold. The threshold value can be the percentage of areas for which

- 52 042157 the presence of a deletion or amplification is determined.

The specified threshold value for the number of areas may also depend on how pronounced the imbalance is for the counted areas. For example, the number of regions used as a threshold for determining cancer classification may depend on the specificity and sensitivity (aberrant threshold) for detecting aberration in each region. For example, if the aberrant threshold is low (eg, Z-score is 2), a high threshold number (eg, 150) may be selected. However, if the aberrant threshold is high (eg, Z-score is 3), the threshold number may be lower (eg, 50). The number of areas showing aberration can also be a weighted value, for example, one area that shows a high imbalance can be given more weight than an area that shows only a slight imbalance (i.e. there are more classifications than just positive and negative aberration results). For example, the sum of Z-scores can be used, thus applying weighted values.

Accordingly, the number (which may include the number and/or size) of chromosomal regions showing significant over- or under-representation of the estimated normalization marks (or other appropriate value for a group property) can be used to reflect disease severity. The number of chromosomal regions with an aberrant count of normalization marks can be determined by two factors, namely the number (or size) of chromosomal aberrations in tumor tissues and the fractional concentration of tumor-derived DNA in a biological sample (eg, plasma). With more advanced cancers, there is a tendency for more (and more) chromosomal aberrations. Accordingly, more cancer-associated chromosomal aberrations can potentially be detected in a sample (eg, plasma). In patients with more advanced cancer, a higher tumor burden will result in a higher fractional plasma concentration of tumor-derived DNA. As a result, tumor-associated chromosomal aberrations in the plasma sample are easier to detect.

One possible way to increase sensitivity without sacrificing specificity is to take into account the result for an adjacent chromosome segment. In one embodiment, the Z-score cutoff points remain at >2 and <-2. However, a chromosomal region is only classified as potentially aberrant if the same type of aberration occurs in two consecutive segments, eg both segments have a Z-score >2. In other embodiments, the Z-scores of neighboring segments may be summed using a higher limit value. For example, the Z-scores of three consecutive segments may be summed, and a limit value of 5 may be applied. This principle may be extended to more than three consecutive segments.

The combination of quantitative and aberrant thresholds may also depend on the purpose of the analysis, and any prior data on the organism (or lack thereof). For example, when screening a normal healthy population for cancer, high specificity is typically used, potentially both in terms of the number of regions (i.e., a high threshold for the number of regions) and the aberrant threshold in identifying a region as containing an aberration. However, for a patient at higher risk (for example, a patient with a tumor complaint or a patient with a family history, a smoker, a chronic human papillomavirus (HPV) infection carrier, a hepatitis virus carrier, or another virus carrier), the thresholds may be lower to provide more sensitivity (fewer false negatives).

In one embodiment, when using 1 Mb resolution for chromosome aberration detection and a lower detection limit of 6.3% tumor-derived DNA, the number of molecules in each 1 Mb segment would be 60,000. When translated, this would give approximately 180 million (60 000 reads/Mb x 3000 Mb) aligned reads for the whole genome.

Smaller segment size provides higher resolution for detecting smaller chromosomal aberrations. However, this increases the total number of molecules required for analysis. A larger segment size reduces the number of molecules required for analysis at the expense of resolution. Accordingly, only larger aberrations can be detected. In one embodiment, larger regions may be used, segments exhibiting aberration may be separated and the resulting sub-regions analyzed for better resolution (eg, as described above). With estimated data for the size of the deletion or amplification to be detected (or the minimum concentration for detection), the number of molecules for analysis can be determined.

B. CNA based on sequencing of bisulfite-treated plasma DNA.

Genome-wide hypomethylation and CNA are frequently observed in tumor tissues. In this description, the authors of the present invention demonstrate that information about CNA and cancer-related changes in methylation can be simultaneously obtained by bisulfite sequencing of plasma DNA. Since these two types of analysis can be performed on the same dataset, there is little to no additional cost for CNA analysis. According to others

- 53 042157 embodiments, separate procedures can be carried out to obtain information about methylation and genetic information. In other embodiments, a similar cancer-associated hypermethylation assay may be performed in combination with a CNA assay.

In FIG. 37A shows CNA analysis for tumor tissues, on bisulfite-free (BS) plasma DNA and bisulfite-treated plasma DNA (from center to edges), for patient TBR36. In FIG. 37A shows CNA analysis for tumor tissues, on bisulfite-free (BS) plasma DNA and bisulfite-treated plasma DNA (from center to edges), for patient TBR36. The outer circle represents the ideogram of the chromosome. Each dot represents a result for a 1MB area. Green, red, and gray dots indicate areas with copy addition, copy loss, and no change in the number of copies, respectively. Z-scores for plasma analysis are given. The difference between the two concentric lines is 5. The copy number for tumor tissue analysis is shown. The difference between the two concentric lines is 1 copy. In FIG. 38A shows CNA analysis for tumor tissues, on bisulfite-untreated (BS) plasma DNA and bisulfite-treated plasma DNA (from center to edges), for patient TBR34. The CNA patterns detected in bisulfite-treated and non-bisulfite-treated plasma samples were in agreement.

The CNA patterns detected in tumor tissues, non-bisulfite-treated plasma and bisulfite-treated plasma were consistent. To further evaluate the agreement between the results for bisulfite-treated and non-bisulfite-treated plasma, a scatter plot was plotted. Fig. 37B is a scatterplot showing the relationship between Z-scores for CNA detection in 1 Mb patches using bisulfite-treated and non-bisulfite-treated plasma for patient TBR36. There was a positive correlation between the Z-scores of these two analyzes (r=0.89, p<0.001, Pearson's correlation). Fig. 38B is a scatterplot showing the relationship between Z-scores for CNA detection in 1 Mb patches using bisulfite-treated and non-bisulfite-treated plasma for patient TBR34. There was a positive correlation between the Z-scores of these two analyzes (r=0.81, p<0.001, Pearson's correlation).

C. Synergistic analysis of cancer-associated CNAs and methylation changes.

As described above, a CNA assay may include counting the number of sequencing reads in each 1 Mb region, and a methylation density assay may include the proportion of methylated cytosine residues in CpG dinucleotides. Combining these two assays may provide synergistic information for cancer detection. For example, the methylation classification and the CNA classification can be used to determine a third cancer level classification.

In one embodiment, the presence of either cancer-associated CNAs or methylation changes can be used to establish the potential presence of a cancer. In such an embodiment, the sensitivity of cancer detection can be increased in the presence of either CNA or methylation changes in the plasma of the test subject. According to another implementation variant, the presence of both changes can be used to determine the presence of cancer. According to such an embodiment, the specificity of the test can be increased, since changes in either of these two types can potentially be detected in some non-cancer subjects. Accordingly, the third classification can only be positive for cancer if both the first classification and the second classification indicate cancer.

The study included 26 patients with HCC and 22 healthy subjects. A blood sample was taken from each subject and the plasma DNA was sequenced after treatment with bisulfite. In patients with HCC, blood samples were taken at the time of diagnosis. The presence of a significant amount of CNA was defined, for example, as the presence of >5% of areas showing a Z-score of <-3 or >3. The presence of significant cancer-associated hypomethylation was defined as the presence of >3% of regions showing a Z-score of <-3. The number of areas (sections) can be represented, for example, as the number of sections calculated without processing, the percentage and length of the indicated sections.

In table. 3 shows the detection of significant amounts of CNA and changes in plasma methylation in 26 HCC patients using massively parallel sequencing of bisulfite-treated plasma DNA.

Detection rates for cancer-related changes in methylation and CNA were and 50%, respectively. Detection rate (i.e. diagnostic sensitivity) increased to 73%

- 54 042157 when using the presence of any indicator to establish the potential presence of cancer.

The results for two patients are shown, reflecting either the presence of CNA (FIG. 39A) or changes in methylation (FIG. 39B). Fig. 39A is a Circos plot representing CNA (inner circle) and methylation analysis (outer circle) for bisulfite-treated plasma from a TBR240 HCC patient. In the CNA analysis, green, red, and gray dots indicate areas with additions of chromosomal material, losses of chromosomal material, and areas with no change in the number of copies, respectively. In the methylation analysis, green, red, and gray dots indicate areas of hypermethylation, hypomethylation, and normal methylation, respectively. In said patient, cancer-associated CNAs were detected in plasma, while methylation analysis did not detect significant cancer-associated hypomethylation. Fig. 39B is a Circos plot representing CNA (inner circle) and methylation analysis (outer circle) for bisulfite-treated plasma from a TBR164 HCC patient. Plasma hypomethylation associated with cancer was detected in this patient. However, no significant amounts of CNA were observed. These two patient results, reflecting the presence of both CNA and methylation changes, are shown in FIG. 48A (TBR36) and 49A (TBR34).

In table. 4 shows the detection of significant amounts of CNA and methylation changes in the plasma of 22 control subjects using massively parallel sequencing of bisulfite-treated plasma DNA. Bootstrapping (i.e., one sample exclusion) was used to evaluate each of the control subjects. Accordingly, when evaluating a particular subject, the remaining subjects (21 subjects) were used to calculate the mean and SD for the control group.

Table 4

CNA _______________________ Presence Absence

Change in methylation Availability 1 2 Absence 1 18

The specificity for detecting significant changes in methylation and CNA was 86% and 91%, respectively. Specificity increased to 95% if both were required to be present to establish the potential presence of cancer.

In one embodiment, samples that are positive for CNA and/or hypomethylation are considered positive for cancer, and samples that detect neither are considered negative. The use of a logic circuit or provides higher sensitivity. In another embodiment, only samples that are positive for both CNA and hypomethylation are considered positive for cancer, thus providing higher specificity. According to another implementation variant, three levels of classification can be used. Subjects are classified by having: i) both normal values; ii) one abnormal indicator; sh. both anomalies.

For these three classifications, different follow-up strategies may be applied. For example, for subjects (iii), the most intensive follow-up protocol may be used, eg including whole body imaging; for subjects (ii) a less intensive follow-up protocol may be used, eg repeated plasma DNA sequencings over a relatively short period of time of a few weeks; and for subjects (i) the least intensive follow-up protocol may be used, such as retesting after a few years. In other embodiments, the measurement of methylation and CNA can be used in combination with other clinical parameters (eg, imaging results or serum chemistry) to further refine the classification.

D. Predictive value of plasma DNA analysis after curative treatment.

The presence of cancer-associated CNAs and/or plasma methylation changes indicates the presence of tumor-derived DNA in the bloodstream of a cancer patient. After treatment (eg, surgery), a decrease or disappearance of these cancer-related changes is expected. On the other hand, persistence of these changes in plasma after treatment may indicate incomplete removal of all tumor cells from the body and may be suitable for predicting disease recurrence.

Blood samples were taken from two patients with HCC, TBR34 and TBR36, one week after curative surgical removal of tumors. CNA and methylation analysis was performed on bisulfite-treated postoperative plasma samples.

In FIG. 40A shows CNA analysis on bisulfite-treated plasma DNA collected before (inner circle) and after (outer circle) surgical removal of a tumor in a patient with HCC, TBR36. Each dot represents a result for a 1 MB area. Green, red, and gray dots indicate areas with increased copy count, lost copies, and no change in copy count, respectively. Most of the CNA observed before treatment disappeared after tumor removal. Share of plots,

- 55 042157 showing Z-score <-3 or >3, decreased from 25% to 6.6%.

In FIG. 40B shows methylation analysis of bisulfite-treated DNA plasma collected before (inner circle) and after (outer circle) surgical tumor removal in a TBR36 HCC patient. Green, red, and gray dots indicate areas of hypermethylation, hypomethylation, and normal methylation, respectively. There was a marked decrease in the proportion of sites showing significant hypomethylation from 90% to 7.9%, and there was also a marked decrease in the degree of hypomethylation. This patient achieved complete clinical remission 22 months after tumor removal.

In FIG. 41A shows CNA analysis of bisulfite-treated plasma DNA collected before (inner circle) and after (outer circle) surgical removal of a tumor in a patient with HCC, TBR34. Although there is a decrease in both the number of sites showing CNA and the magnitude of CNA in the affected areas after surgical removal of the tumor, residual CNA can be observed in the postoperative plasma sample. The red circle indicates the area where residual CNA is most evident. The proportion of sites showing a Z-score <-3 or >3 decreased from 57 to 12%.

In FIG. 41B. shows methylation analysis of bisulfite-treated plasma DNA collected before (inner circle) and after (outer circle) surgical removal of a tumor in a TBR34 HCC patient. The amount of hypomethylation decreased after tumor removal, with a decrease in the average Z-score for hypomethylated areas from -7.9 to -4.0. However, for the proportion of plots with a Z-score <-3, the opposite change was observed, with an increase from 41% to 85%. This observation potentially indicates the presence of residual cancer cells after treatment. Clinically, several foci of tumor nodes were found in the unresectable liver 3 months after tumor removal. From the 4th month after surgery, metastases to the lungs were observed. This patient died as a result of local recurrence and metastasis 8 months after surgery.

Observations in these two patients (TBR34 and TBR36) suggest that the presence of residual cancer-associated CNA changes and hypomethylation can be used for monitoring and prognosis in cancer patients after curative treatment. The data also showed that the rate of change in plasma CNAs detected can be used synergistically with the rate of change in plasma DNA hypomethylation to predict and monitor treatment efficacy.

Accordingly, in some embodiments, one biological sample is obtained prior to treatment and a second biological sample is obtained after treatment (eg, surgery). For the first sample, first values are obtained, eg, Z-scores for regions (eg, region methylation levels and normalized CNA values), and the number of regions showing hypomethylation and CNA (eg, amplification or deletion). For the second sample, second values may be obtained. According to another implementation variant, after treatment, a third or even subsequent additional samples can be obtained. The number of regions showing hypomethylation and CNA (eg, amplification or deletion) can be obtained for the third or even subsequent additional samples.

As described above for FIG. 40A and 41A, the first number of areas showing hypomethylation for the first sample can be compared with the second number of areas showing hypomethylation for the second sample. As described above for FIG. 40B and 41B, the first number of regions showing hypomethylation for the first sample can be compared with the second number of regions showing hypomethylation for the second sample. Comparison of said first number with said second number, and the first number with the second number may be used to determine treatment prognosis. In various embodiments, only one of the comparisons may be indicative of the prediction, or both comparisons may be used. In embodiments where a third or subsequent additional samples are obtained, one or more of these samples may be used to determine treatment prognosis, either alone or in combination with a second sample.

In one embodiment, a worse prediction is made if the first difference between the first amount and the second amount is below a threshold for the first difference. In another embodiment, a worse prediction is made if the second difference between the first number and the second number is below a threshold for the second difference. The threshold can be the same or different. In one embodiment, the threshold for the first difference and the threshold for the second difference are 0. Accordingly, in the example above, the difference in methylation values indicates a worse prognosis for a TBR34 patient.

The prediction may be better if the first difference and/or the second difference are above the same threshold or respective thresholds. The classification for the prediction may depend on how much above or below the threshold the differences are. Several thresholds can be used to obtain different classifications. Larger differences can predict better outcomes, while smaller differences (and even negative values) can predict worse outcomes.

- 56 042157

In some embodiments, the times when different samples are taken may also be taken into account. With such time parameters, it is possible to determine the kinetics or rate of change of the quantity. In one embodiment, a rapid decrease in plasma tumor-associated hypomethylation and/or a rapid decrease in plasma tumor-associated CNAs indicates a good prognosis. Conversely, a static or rapid increase in tumor-associated plasma hypomethylation and/or a static or rapid increase in tumor-associated CNAs indicates a poor prognosis. Methylation and CNA measurements can be used in combination with other clinical parameters (eg, imaging findings, serum chemistry, or protein markers) to predict clinical outcome.

In embodiments, samples other than plasma may be used. For example, tumor-associated methylation aberrations (eg, hypomethylation) and/or tumor-associated CNAs can be measured in tumor cells circulating in the blood of cancer patients, cell-free DNA, or tumor cells in urine, stool, saliva, sputum. , biliary fluid, pancreatic fluid, cervical smears, secretions of the genital tract (eg, vagina), ascitic fluid, pleural fluid, seminal fluid, sweat and tears.

In various embodiments, tumor-associated methylation aberrations (e.g., hypomethylation) and/or tumor-associated CNAs can be detected in the blood or plasma of patients with breast cancer, lung cancer, colon and rectal cancer, pancreatic cancer, ovarian cancer, nasopharyngeal carcinoma, cervical cancer, melanoma, brain tumors, etc. Thus, since methylation and genetic changes such as CNA are a universal phenomenon in cancers, the described methods can be applied to all types of cancer. Methylation and CNA measurements can be used in conjunction with other clinical parameters (eg, imaging findings) to predict clinical outcome. Embodiments may also be used to screen and monitor patients with precancerous conditions such as adenomas.

Accordingly, in one embodiment, said biological sample is taken before treatment, and CNA and methylation measurements are repeated after treatment. These measurements may give the next first number of regions for which the presence of a deletion or amplification has been determined, and may give the next second number of regions that have been determined to have a methylation level of the region in excess of the corresponding cut-off value for the region. The first number may be compared with the next first number, and the second number may be compared with the next second number to determine the prognosis for the specified organism.

The comparison to determine a prediction for said organism may include determining a first difference between the first count and the next first count, and said first difference may be compared to one or more thresholds for the first difference to determine the prediction. The comparison to determine a prediction for said organism may also include determining a second difference between the second amount and the next second amount, and said second difference may be compared to one or more thresholds for the second difference. The threshold value can be 0 or another number.

The prediction may be worse if the first difference is below the threshold for the first difference, compared to the prediction if the first difference is above the threshold for the first difference. The prediction may be worse if the second difference is below the threshold for the second difference compared to the prediction if the second difference is above the threshold for the second difference. Examples of treatments include immunotherapy, surgery, radiation therapy, chemotherapy, antibody-based therapy, gene therapy, epigenetic therapy, or targeted therapy.

E. Efficiency.

The diagnostic performance of CNA and methylation analysis for different numbers of sequencing reads and regions of different sizes is described below.

1. Number of sequencing reads.

In accordance with one embodiment, the authors of the present invention examined the plasma DNA of 32 healthy controls, 26 patients suffering from hepatocellular carcinoma and 20 patients suffering from other types of cancers, including nasopharyngeal carcinoma, breast cancer, lung cancer, neuroendocrine cancer and smooth muscle sarcoma. 22 out of 32 healthy subjects were randomly selected as the reference group. The mean and standard deviation (SD) of these 22 reference individuals were used to determine the normal range for methylation density and genome representation. DNA extracted from each individual's plasma sample was used to prepare a sequencing library using the Illumina Paired End Sequencing Kit. The sequencing libraries were then treated with bisulfite, which converts unmethylated cytosine residues to uracil. Bisulfite-converted sequencing library for each plasma sample

- 57 042157 sequenced on one lane of the Illumina HiSeq2000 sequencer.

After base recognition, adapter sequences and low quality bases (ie, quality score <5) at the ends of the fragments were removed. The FASTQ clipped reads were then processed using the Methy-Pipe methylation data analysis system (P Jiang et al. 2010, IEEE International Conference on Biomformatics and Biomedicine, CIO:10.1109/BIBMW.2010.5703866). To align bisulfite-converted sequencing reads, the present inventors first performed in silico conversion of all cytosine to thymines residues on Watson-Crick strands individually using a reference human genome (NCBI version 36/hg19). The present inventors then performed an in silico conversion of each cytosine to thymine in all processed reads and stored information about the location of each converted residue. SOAP2 was used to align converted reads to two pre-transformed human reference genomes (R Li et al. 2009 Bioinformatics 25:1966-1967), with a maximum of two mismatches for each read aligned. Only reads mapping with unique genomic localization were used for further analysis. Ambiguous reads that mapped to both Watson-Crick strands and duplicate (clonal) reads were removed. Cytosine residues in the context of CpG dinucleotides were used for subsequent methylation analysis. After alignment, cytosines originally present in the sequencing reads were reconstructed based on the location information preserved by in silico conversion. Reduced cytosines in CpG dinucleotides were counted as methylated. Thymines in CpG dinucleotides were counted as unmethylated.

To analyze methylation, the genome was divided into regions of equal size. Tested chunk sizes include 50kb, 100kb, 200kb, and 1Mb. The methylation density for each site was calculated as the number of methylated cytosines in the context of CpG dinucleotides divided by the total number of cytosines at CpG positions. In other embodiments, the size of the regions may not be the same throughout the genome. In one embodiment, each of these unequally sized patches is compared across multiple subjects.

To determine whether the plasma methylation density in the test case is normal, the methylation density was compared with the results of the reference group. Twenty-two of 32 healthy subjects were randomly selected as a reference group for the calculation of the methylation Z-score (Z _meth ).

MD _lest - MD ref MD _sd

~

Where is the methylation density in the test case for a specific area of 1 Mb;

^Dref - the average value of the methylation density of the reference group for the corresponding site; and - SD methylation density of the reference group for the respective site.

For CNA analysis, the number of sequencing reads mapped to each 1 Mb region was determined (KCA Chan et al. 2013 Clin Chem 59:211-24). Sequencing read density was determined for each region after correcting for GC bias using locally weighted regression to smooth the scatterplot as described (EZ Chen et al. 2011 PLoS One 6: e21791). For plasma analysis, the density of sequencing reads in the test case was compared with the reference group to calculate the CNA Z-score (au):

. RDlesl ~ RD ref ^CNA RD _sd

RD

Where is the density of sequencing reads for the test case for a specific 1 Mb area;

RD _r rf is the mean density of sequencing reads for the reference group for the corresponding RDc^ SP 1 of the site; and ~ density of sequencing reads for the reference group for the respective region. A site was defined as showing CNA if the Zcna of said site was <3 or >3.

On a per-case basis, a median leveled read of 93M (range: 39M-142M) was obtained. To evaluate the effect of reducing the number of sequencing reads on diagnostic performance, the present inventors randomly selected 10 million alignment reads in each case. The same set of reference individuals were used to establish a reference range for each 1 Mb region for the reduced sequencing read data set. For each case, the percentage of sites showing significant hypomethylation, i.e., was determined. Z _met h < -3, and the percentage of sites with CNA, i.e. Zcna <-3 or >3. Receiver Functional Characteristics (ROC) curves were used to illustrate the diagnostic performance of genome-wide hypomethylation and CNA analysis for datasets with

- 58 042157 all sequencing reads from 1 lane and 10 million reads per case. All healthy subjects (32 subjects) were used in the ROC analysis.

In FIG. 42 shows a diagram of the diagnostic performance of genome-wide hypomethylation analysis with different numbers of sequencing reads. For the analysis of hypomethylation, the area under the curve for the ROC curves did not differ significantly between the two data sets that examined all sequencing reads per lane and 10 million reads per case (P=0.761). The diagnostic performance of the CNA assay deteriorated, with a significant decrease in area under the curve, as the number of sequencing reads decreased from single lane data to 10 million (P<0.001).

2. The effect of applying areas of various sizes.

In addition to dividing the genome into 1 Mb regions, the present inventors also explored the possibility of using smaller regions.

Theoretically, the use of smaller patches could potentially reduce the variability in methylation density within a patch. This is due to the fact that the methylation density of different genomic regions can vary widely. If the site size is larger, the probabilities of including regions with different methylation densities increase and, accordingly, lead to an overall increase in the variability in the methylation density of the sites.

Although the use of smaller regions can potentially reduce the variability in methylation density associated with differences between regions, on the other hand, this will reduce the number of sequencing reads mapped to a particular region. With a decrease in the mapping of reads to individual sites, variability will increase due to sample variations. The optimal region size that can provide the minimum overall methylation density variability can be established experimentally based on the requirements of a particular diagnostic application, for example, taking into account the total number of sequencing reads per sample and the type of DNA sequencer used.

In FIG. 43 is a graph showing ROC curves for cancer detection based on a genome-wide analysis of hypomethylation at different patch sizes (50 kb, 100 kb, 200 kb and 1 Mb). P-values are shown to compare area under the curve at 1 Mb patch size. There is a tendency to improve with a decrease in the size of the site from 1 Mb to 200 kb.

F. Total Probability Index.

The number of methylation regions and CNAs can take on different values. The examples above describe the number of regions above a cut-off or the percentage of such regions showing significant hypomethylation or CNA as a parameter for classifying a sample as associated with cancer. Such methods do not take into account the amount of aberration for individual sections. For example, a site with a Zmeth of -3.5 would be classified the same as a site with a Zmeth of -30, since they would both be classified as significantly hypomethylated. However, the degree of changes in plasma hypomethylation, i.e. The magnitude of the Zmeth value is affected by the amount of cancer-associated DNA in the sample and, accordingly, it can complement the information on the percentage of areas showing aberrations, reflecting the tumor load. A higher fractional concentration of tumor DNA in a plasma sample will result in a lower methylation density and this will result in a lower Z _meth value.

1. The indicator of the total probability as a diagnostic parameter.

To use information about the amount of aberrations, the authors of the present invention have developed a method called the indicator of the total probability (CP). Based on the normal distribution of the probability function, each Z _meth value was translated into the probability of such an observation occurring by chance.

The SR indicator was calculated as:

CP score = -log(Prob.) for site (i) with Zmeth<-3 where Probi is the probability Z _meth for site (i) according to Student's t-distribution with 3 degrees of freedom, and log is a function of the natural logarithm . According to another implementation variant, the logarithm to base 10 (or other number) may be used. In other embodiments, other distributions may be used, such as but not limited to the listed normal distribution and gamma distribution, to convert Z-score to SR.

A higher CP indicates a lower chance of such a deviation in methylation density occurring by chance in a normal population. Thus, a high CP would indicate a higher likelihood of abnormal DNA hypomethylation in the sample, such as the presence of cancer-associated DNA.

Compared to the percentage of areas showing aberrations, the measured CP has a higher dynamic range. Although tumor burden in different patients may vary

- 59 042157 vary over a wide range, a larger range of CP values may be appropriate to reflect tumor burden in patients with relatively high and relatively low tumor burden. In addition, the use of SR measures could potentially provide more sensitive detection of changes in plasma concentrations of tumor-associated DNA. This provides benefits in monitoring response to treatment and prognosis. Accordingly, a decrease in CP values during treatment indicates a good response to treatment. The absence of a decrease or even an increase in CP during treatment indicates an unsatisfactory or absent response. When predicting, a high CP score indicates a high tumor burden and suggests a poor prognosis (eg, higher chance of death or tumor progression).

In FIG. 44A shows the diagnostic performance of the sum probability (CP) and percentage of areas with aberrations. There was no significant difference between the area under the curve for the two types of diagnostic algorithm (P=0.791).

In FIG. 44B shows the diagnostic performance of plasma assays for global hypomethylation, CpG island hypermethylation, and CNA. At one sequencing lane per sample (200 kb patch size for hypomethylation assay and 1 Mb patch size for CNA; CpG islands defined according to the University of California, Santa Cruz (UCSC) database), area under curve values for all three assay types were greater than 0 .90.

In all subsequent analyses, the maximum CP for control subjects was used as the cut-off point for each of the three types of analysis. The choice of these cut-off points provided a diagnostic specificity of 100%. Diagnostic sensitivity for the analysis of total hypomethylation, hypermethylation of CpG islands and CNA was 78%, 89% and 52%, respectively. In 43 out of 46 cancer patients, at least one of the three types of aberrations were detected, resulting in a sensitivity of 93.4% and a specificity of 100%, respectively. Our results show that these three assays can be used synergistically for cancer detection.

In FIG. 45 is a table of the results of the analysis of global hypomethylation, hypermethylation of CpG islands and CNA in patients with hepatocellular carcinoma. The cut-off values for the CP index for the three types of analysis were 960, 2.9 and 211, respectively. Positive results for the CP score are in bold and underlined.

In FIG. 46 is a table showing the results of global hypomethylation, CpG island hypermethylation, and CNA analysis in patients with cancers other than hepatocellular carcinoma. The cutoff values for the CP index for the three types of analysis were 960, 2.9, and 211, respectively. Positive results for the CP score are in bold and underlined.

2. Application of the SR indicator for cancer monitoring.

Serial samples were obtained from a patient with HCC TBR34 before and after treatment. These samples were tested for global hypomethylation.

In FIG. 47 shows a serial analysis of plasma methylation for patient TBR34. The inner circle reflects the methylation density of the buffy coat (black) and tumor tissues (purple). For plasma samples, Z _meth is shown for each 1 Mb patch. The difference between the two lines reflects the Z _meth difference of 5. Red and gray dots indicate areas with hypomethylation and no change in methylation density compared to the reference group. Starting from the second from the center of the circle in the direction of the outer circle, plasma samples obtained before treatment, 3 days and 2 months after tumor removal, respectively, are presented. Prior to treatment, a high degree of hypomethylation could be observed in plasma, and more than 18.5% of the sites had Z _meth , <-10. 3 days after tumor removal, a decrease in the degree of plasma hypomethylation can be observed in the absence of C Z _meth <-10 sites.

Table 5

Case No. Moment of time Methylation analysis Percentage of sites showing significant hypomethylation Total Probability Index (CP) Summative Z-Score TBR34 Before surgery (OT) 62.6% 37.573 14.285 3 days after OT 80.5% 17.777 9.195 2 months after OT 40.1% 15.087 5.201

Tab. 5 demonstrates that despite a decrease in the magnitude of hypomethylation changes at 3 for after surgical removal of the tumor, the percentage of sites where aberrations were observed paradoxically increased. On the other hand, CP was more accurate in detecting a reduction in plasma hypomethylation and may better reflect changes in tumor burden.

At 2 months post OT, a significant percentage of sites were still present, demonstrating

- 60 042157 changes in hypomethylation. The HR also remained static at approximately 15,000. Later, after 3 months, this patient was diagnosed with multifocal tumor metastases (previously unknown at the time of surgery) in the unresectable part of the liver, and several lung metastases were found 4 months after operations. The patient died as a result of a metastatic process 8 months after the operation. These results suggest that the CP score may be more effective than the percentage of aberrations in reflecting tumor burden.

In general, CP may be suitable for use in cases where measurement of the amount of tumor DNA in plasma is required. Examples of such applications include: predicting and monitoring cancer patients (eg, monitoring response to treatment, or monitoring tumor progression).

The summative Z-score is the direct sum of the Z-scores, i.e. without converting to probability. In this example, the summed Z-score exhibits the same behavior as the CP score. In other cases, SR may be more sensitive than the summed Z-score for monitoring residual disease due to the greater dynamic range of the CP score.

X. Effect of CNA on methylation

The use of CNA and methylation to determine the respective classifications of the level of cancer, with the combination of these classifications to obtain a third classification, is described above. In addition to such pooling, CNA can be used to modify cutoffs for methylation assays and to identify false positives by comparing methylation levels for groups of regions with different CNA characteristics. For example, the level of methylation at excess content (eg, Z _CNA >3) can be compared with the level of methylation at normal content (eg, -3<Z _CNA <3). First, the effect of CNA on methylation levels is described.

A. Change in methylation density in areas with addition or loss of chromosomal material.

Since general hypomethylation is commonly observed in tumor tissues, the presence of tumor-derived DNA in the plasma of cancer patients will result in a decrease in methylation density compared to non-cancer subjects. Theoretically, the degree of hypomethylation in the plasma of cancer patients is proportional to the fractional concentration of tumor-derived DNA in the plasma sample.

In the case of regions for which there is an addition of chromosomal material in tumor tissues, an additional dose of tumor DNA is released from the amplified DNA segments into the plasma. This increased contribution of tumor DNA to plasma theoretically results in a higher degree of hypomethylation of the affected area of plasma DNA. An additional factor is that the genomic regions for which amplification is observed are expected to stimulate the growth of tumor cells, and, accordingly, are expected to be expressed. Such areas are usually hypomethylated.

In contrast, for areas where there is a loss of chromosomal material in the tumor tissue, a reduced contribution of tumor DNA to plasma will result in a lower degree of hypomethylation compared to areas with no change in copy number. An additional factor is that genomic regions that are deleted in tumor cells may contain tumor suppressor genes, and silencing of such regions may be beneficial for tumor cells. Accordingly, the likelihood of hypermethylation of such regions would be expected to be higher.

In the present invention, the authors use the results of two patients with HCC (TBR34 and TBR36) to illustrate this effect. In FIG. 48A (TBR36) and 49A (TBR34) circles highlight areas with addition or loss of chromosomal material and present the corresponding methylation analysis. Fig. 48B and 49B are plots of methylation z-scores at loss, normal, and addition for TBR36 and TBR34 patients, respectively.

In FIG. 48A is a Circos plot showing CNA (inner circle) and methylation changes (outer circle) in bisulfite-treated plasma DNA from a TBR36 HCC patient. Areas with addition or loss of chromosomal material are marked with red circles. Areas where chromosomal material was added were more hypomethylated than areas without copy number changes. Areas where there is loss of chromosomal material were less hypomethylated than areas with no change in copy number. Fig. 48B is a plot of methylation Z-scores for regions with addition or loss of chromosomal material and regions without change in copy number for a TBR36 HCC patient. Compared to regions with no copy number changes, regions with addition of chromosomal material had more negative Z-scores (more hypomethylation), and regions with loss of chromosomal material had less negative Z-scores (less hypomethylation).

In FIG. 49A is a Circos plot showing CNA (inner circle) and methylation changes (outer circle) in bisulfite-treated plasma DNA from a TBR34 HCC patient. Fig. 49B

- 61 042157 is a plot of methylation Z-scores for areas with addition or loss of chromosomal material, and areas without change in copy number for a TBR34 HCC patient. The difference in methylation density between areas with addition or loss of chromosomal material was greater in the TBR36 patient compared to the TBR34 patient, since the fractional concentration of tumor-derived DNA in the first patient was higher.

In this example, the same regions are used to determine the CNA as for the determination of methylation. According to one implementation variant, the respective limit values for the regions depend on whether a deletion or amplification is observed in the corresponding region. In one embodiment, the value of the corresponding cut-off value for a region (e.g., the Z-score cutoff point used to determine hypomethylation) is greater when amplification is observed in the corresponding region than when no amplification is observed (e.g., the indicated value may exceed 3, and a cutoff point <-3 may apply). Accordingly, when testing for hypomethylation, the respective cut-off value for a region may have a larger negative value in cases where amplification is observed in the respective region than in cases where no amplification is observed. It is expected that such an implementation option will increase the specificity of testing for the detection of cancers.

According to another implementation variant, the value of the corresponding limit value for the region is less (eg, less than 3) in cases where a deletion is observed in the corresponding region than in cases where no deletion is observed. Accordingly, when testing for hypomethylation, the corresponding cut-off value for a region may be less negative in cases where a deletion is observed in the relevant region than in cases where no deletion is observed. It is expected that such an implementation option will increase the sensitivity of testing for the detection of cancers. The limit value correction according to the embodiments described above can be changed depending on the required sensitivity and specificity for a particular diagnostic scenario. In other embodiments, methylation and CNA measurements can be used in conjunction with other clinical parameters (eg, imaging results or serum chemistry) to predict cancers.

B. Using CNA to select areas.

As described above, the present inventors have shown that plasma methylation density changes in areas containing copy number aberrations in tumor tissues. In areas with increased copy number in tumor tissue, an increased contribution of hypomethylated tumor DNA to plasma will result in a greater degree of plasma DNA hypomethylation compared to areas without copy number aberrations. Conversely, in areas of copy loss in tumor tissue, a reduced plasma contribution of hypomethylated cancer-derived DNA will result in less hypomethylation of plasma DNA. This relationship between plasma DNA methylation density and relative abundance can potentially be used to differentiate hypomethylation outcomes associated with the presence of cancer-associated DNA and other non-cancer causes (eg, SLE) of plasma DNA hypomethylation.

To illustrate this method, the authors of the present invention examined plasma samples of two patients with hepatocellular carcinoma (HCC) and two patients with SLE without cancer. In these two patients with SLE (SLE04 and SLE10), there was an obvious presence of hypomethylation and CNA in plasma. The SLE04 patient had hypomethylation in 84% of the sites and CNA in 11.2% of the sites. The SLE10 patient had hypomethylation in 10.3% of the sites and CNA in 5.7% of the sites.

In FIG. 50A and B show the results of plasma hypomethylation and CNA analysis for patients with SLE, SLE04, and SLE10. The outer circle displays methylation Z-scores (Z _meth ) at 1 MB resolution. Areas with Z-methylation index Z _meth <-3 are marked in red, with Z _meth >-3 - in gray. The inner circle displays CNA Z-scores (Z _CNA ). Green, red, and gray dots indicate areas with Z _CNA >3, <3, and -3 to 3, respectively. These two patients with SLE showed changes in plasma hypomethylation and CNA.

To determine if changes in methylation and CNA were consistent with the presence of cancer-derived DNA in plasma, the present inventors compared Zmeth for regions with Z _CNA >3, <-3 and -3 to 3. With changes in methylation and CNA, in contributed by cancer-derived plasma DNA, regions with Z _CNA <-3 are expected to be less hypomethylated and have less negative Z _meth . In contrast, areas with Z _CNA >3 are expected to be more hypomethylated and have more negative Zmeth. For clarity, the present inventors used a one-tailed rank sum test to compare Zmeth for regions with CNA (i.e., regions with Z _CNA <-3 or >3) with regions without CNA (i.e., regions with Z _CNA from -3 to 3). Other implementations may apply other statistical criteria, for example,

- 62 042157 but not limited to Student's t-test, analysis of variance (ANOVA), and Kruskal-Wallis test.

In FIG. 51A and B are Z _meth analyzes for regions with and without CNA in the plasma of two HCC patients (TBR34 and TBR36). Areas with Z _CNA <-3 and >3 represent areas under and over represented in plasma, respectively. In both TBR34 and TBR36, regions that were underrepresented in plasma (i.e., regions with Z _CNA <-3) had significantly higher Z _meth (P-value <10 ^-5 , one-tailed rank sum test) compared to with regions normally present in plasma (ie regions with Z _CNA ranging from -3 to 3). Normal representation corresponds to that expected for a euploid genome. Over-represented regions in plasma (ie, regions with Z _CNA >3) had significantly lower Z _meth compared to regions normally presented in plasma (P-value <105, one-tailed rank sum test). All these changes were consistent with the presence of hypomethylated tumor DNA in plasma samples.

In FIG. 51C and D are Z _meth analyzes for regions with and without CNA in the plasma of two SLE patients (SLE04 and SLE10). Areas with Z _CNA <-3 and >3 represent areas under and over represented in plasma, respectively. In SLE04, regions that were underrepresented in plasma (i.e., regions with Z _CNA <-3) did not have a significantly higher Z _meth (P-value = 0.99, one-tailed rank sum test) compared to regions that were normal present in plasma (ie, regions with Z _CNA ranging from -3 to 3); and regions overrepresented in plasma (ie, regions with Z _CNA >3) did not have significantly lower Z _meth compared to regions normally represented in plasma (P-value = 0.68, one-tailed rank sum test). These results differ from the expected changes due to the presence of tumor-derived hypomethylated DNA in plasma. Similarly, in SLE10, regions with Z _CNA <-3 did not have a significantly higher Z _meth compared to regions with Z _CNA ranging from -3 to 3 (P-value = 0.99, one-tailed rank sum test).

The reason for the lack of a typical cancer-associated relationship pattern between Z _meth and Z _cna in SLE patients is that in SLE patients, CNAs are not present in the particular cell type that also exhibits hypomethylation. In this case, the observed apparent presence of CNA and hypomethylation is due to the altered circulating DNA size distribution in patients with SLE. Altered size distributions can potentially alter sequencing read densities for different genomic regions, leading to obvious CNAs since the references are derived from healthy subjects. As described in the preceding sections, there is a correlation between the size of a circulating DNA fragment and its methylation density. Thus, an altered size distribution can also lead to aberrant methylation.

Although areas with Z _CNA >3 had slightly lower levels of methylation compared to areas with Z _CNA ranging from -3 to 3, the P-value for this comparison was significantly higher than that observed for the two cancer patients. In one embodiment, a P-value may be used as a parameter to determine the likelihood of a test case having cancer. According to another implementation variant, the difference Z _meth between normally and aberrantly presented areas can be used as a parameter reflecting the likelihood of having cancers. In one embodiment, a group of cancer patients can be used to establish a correlation between Z _meth and Z _CNA , and to determine thresholds for different parameters, which allows you to determine whether the changes are consistent with the presence of hypomethylated DNA derived from cancer in the tested plasma sample.

Accordingly, in one embodiment, a CNA analysis may be performed to determine a first set of regions, all of which exhibit one or more of: deletion, amplification, or normal representation. For example, in the first set of regions, all regions may show a deletion, or all regions may show amplification, or all regions may show normal representation (eg, contain a normal first number of regions, eg, have a normal Z _meth ). For said first set of regions, a methylation level may be determined (eg, the first methylation level according to method 2800 may correspond to the first set of regions).

The CNA analysis identifies a second set of regions, all of which show either deletion, amplification, or normal presentation. In the specified second set of regions, features different from the first set will be observed. For example, if the regions of the first set were normal, the regions of the second set may show deletion or amplification. The second level of methylation can be calculated based on the respective amounts of DNA molecules methylated at the sites of the regions of the second set.

A parameter relating the first methylation level to the second methylation can then be calculated. For example, a difference or ratio may be calculated and compared with a limit value. The indicated difference or ratio can also be processed using the distribution

- 63 042157 probabilities (for example, within the framework of statistical testing) to determine the probability of obtaining a value, and this probability can be compared with a limit value for determining the level of cancer based on methylation levels. Such a cutoff point can be chosen to distinguish between cancer and non-cancer samples (eg, SLE).

According to one implementation variant, the level of methylation can be determined for the first set of regions or for a mixture of regions (namely, a mixture of regions showing amplification, deletion, and normal). This methylation level can then be compared to the first cut-off point within the first step of the analysis. When the cut-off point is exceeded, which indicates the likelihood of a cancer, the above analysis can be performed to determine if the indication is a false positive. The resulting cancer level classification may accordingly include a comparison of the parameter for the two methylation levels with a second cut-off point.

The first methylation level may be a statistic (eg, mean or median) of the region methylation levels calculated for each region of the first set of regions. The second methylation level may also be a statistic of region methylation levels calculated for each region of the second set of regions. For example, a statistic may be determined using a one-tailed rank sum test, Student's t-test, analysis of variance (ANOVA), or Kruskal-Wallis test.

XI. Classification of types of cancer

In addition to determining the presence or absence of cancer in the body, embodiments allow identification of the type of cancer associated with a sample. Said identification of the type of cancer may involve patterns of global hypomethylation, hypermethylation of CpG islands and/or CNA. These patterns may include clustering patients with a known diagnosis using measured methylation levels of regions, corresponding CNA values for regions, and methylation levels of CpG islands. The results below demonstrate that for organisms with the same type of cancer, similar values are observed for regions and CpG islands; also in patients without cancer, similar values are observed. When clustered, each of the values for an area or island can represent a separate dimension when clustered.

Similar genetic and epigenetic changes are known to occur in the same type of cancer (E Gebhart et al. 2004 Cytogenet Genome Res; 104: 352-358; PA Jones et al. 2007 Cell; 128: 683-692). The present inventors describe below how CNA patterns and methylation changes detected in plasma can be used to determine the origin or type of cancer. Plasma DNA samples from patients with HCC, patients with non-HCC disease, and healthy controls were classified using, for example, hierarchical cluster analysis. This analysis was performed using, for example, the heatmap.2 function of the R script package (cran.r-project.org/web/packages/gplots/gplots.pdf).

To illustrate the potential of this method, the authors of the present invention used two sets of criteria (group A and group B) as examples to identify suitable features for the classification of plasma samples (see table. 6). In other embodiments, other criteria may be applied to identify these features. Traits used included global CNAs at 1 Mb resolution, global methylation density at 1 Mb resolution, and methylation of CpG islands.

Table 6

Global methylation at 1 Mb resolution Criteria group A Criteria group B Criteria >20 cancer cases with Z-score >3 or <-3 >20 cancer cases with Z-score >2.5 or <-2.5 Number of identified features 584 1911 Signs of CNA Criteria group A Criteria group B Criteria >10 cancer cases with Z-score >3 or <-3 >10 cancer cases with Z-score >2.5 <-2.5 Number of identified features 355 759

- 64 042157

Methylation of CpG islands Criteria group A Criteria group B Criteria >5 cancer cases with density methylation deviating from the mean reference value by 2% in specific CpG islands >1 cancer case with density methylation deviating from the mean reference value by 2% in specific CpG islands Number of identified features 110 191

In the first two examples, the authors of the present invention used to classify all signs of CNA, global methylation at 1 Mb resolution and methylation of CpG islands. In other embodiments, other criteria may be applied, for example, but not limited to, the accuracy of the measure of the trait in plasma of the reference group.

In FIG. 52A shows a hierarchical cluster analysis for plasma samples from HCC patients, patients with non-HCC disease, and healthy controls using all of the 1130 group A traits, including 355 CNAs, 584 global methylation traits at 1 Mb resolution, and methylation status of 110 CpG- islets. The color chart at the top represents the sample groups: green, blue, and red represent healthy subjects, HCC, and patients with non-HCC disease, respectively. In general, in these three groups of subjects, there was a trend towards joint clustering. The vertical axis represents classifying features. Features characterized by similar patterns in different subjects were clustered together. These results suggest that patterns of changes in CpG island methylation, changes in genome-wide methylation at 1 Mb resolution, and plasma CNA can potentially be used to determine the origin of cancer in patients with unknown primary lesions.

In FIG. 52B shows a hierarchical cluster analysis for plasma samples from HCC patients, patients with non-HCC disease, and healthy controls using all 2780 group B traits, including 759 CNAs, 1911 global methylation traits at 1 Mb resolution, and methylation status for 191 CpG- islet. The color chart at the top represents the sample groups: green, blue, and red represent healthy subjects, HCC, and patients with non-HCC disease, respectively. In general, in these three groups of subjects, there was a trend towards joint clustering. The vertical axis represents classifying features. Features characterized by similar patterns in different subjects were clustered together. These results suggest that patterns of methylation changes across different sets of CpG islands, genome-wide methylation changes at 1 Mb resolution, and plasma CNA can be used to determine the origin of cancer in patients with unknown primary lesions. The choice of classifying features can be tailored to specific applications. In addition, the prognosis for the type of cancer may be weighted according to the pre-known probabilities of the different types of cancer in the subjects. For example, patients with chronic viral hepatitis are prone to developing hepatocellular carcinoma, and chronic smokers are prone to developing lung cancer. Accordingly, a weighted probability of cancer type can be calculated using, for example, but not limited to, logistic, multiple, or cluster regression.

In other embodiments, a single feature type may be used for classification analysis. For example, in the following examples, only global methylation at 1 Mb resolution, hypermethylation of CpG islands or CNA at 1 Mb resolution was used for hierarchical clustering. The power of differentiation can be different when using different features. Additional refinement of the parameters of classifying features can potentially improve the accuracy of classification.

In FIG. 53A shows a hierarchical cluster analysis for plasma samples from HCC patients, patients with non-HCC disease, and healthy controls using CpG island methylation trait group A. In general, patients with cancer clustered together, while subjects without cancer fell into another cluster. However, the distinction between patients with HCC and those with non-HCC disease was less pronounced when compared with all three types of features.

In FIG. 53B shows a hierarchical cluster analysis for plasma samples from HCC patients, non-HCC patients, and healthy controls using group A global methylation densities at 1 Mb resolution as classifiers. There was a predominant clustering of patients with HCC and other than

- 65 042157

HCC disease.

In FIG. 54A shows a hierarchical cluster analysis for plasma samples from HCC patients, patients with non-HCC disease, and healthy controls using group A global CNAs at 1 Mb resolution as classifying features. There was a predominant clustering of patients with HCC and non-HCC disease.

In FIG. 54B shows a hierarchical cluster analysis for plasma samples from HCC patients, patients with non-HCC disease, and healthy controls using group B methylation densities of CpG islands as classifiers. A preferential clustering of patients with HCC and non-HCC disease can be observed.

In FIG. 55A shows a hierarchical cluster analysis for plasma samples from HCC patients, non-HCC patients, and healthy controls using group B global methylation densities at 1 Mb resolution as classifying features. A preferential clustering of patients with HCC and non-HCC disease can be observed.

In FIG. 55B shows a hierarchical cluster analysis for plasma samples from HCC patients, patients with non-HCC disease, and healthy controls using group B global CNAs at 1 Mb resolution as classifying features. A preferential clustering of patients with HCC and non-HCC disease can be observed.

These hierarchical clustering results for plasma samples suggest that combinations of different features can potentially be used to identify primary cancer types. Further refinement of criteria selection criteria can potentially further improve the accuracy of the classification.

Accordingly, in one embodiment, if a methylation classification indicates a cancer in an organism, the type of cancer associated with said organism can be identified by comparing the methylation level (e.g., the first methylation level according to method 2800, or the methylation level of any region) with the corresponding a value defined for other organisms (i.e. other organisms of the same type, such as humans). Said corresponding value may refer to the same region or set of sites for which the methylation level was calculated. At least two of these other organisms have been identified as having other types of cancers. For example, the corresponding values may be organized into clusters, with two clusters associated with different types of cancer.

Further, if CNA and methylation are used together to obtain a third cancer level classification, CNA and methylation traits can be compared with corresponding values for other organisms. For example, the first number of regions (eg, in FIG. 36) where a deletion or amplification occurs can be compared with corresponding values determined for other organisms to identify the type of cancer associated with said organism.

In some embodiments, methylation traits are represented by methylation levels of regions from multiple regions of the genome. Regions that have been determined to have region methylation levels in excess of the corresponding region limit may be used, for example, methylation levels of a region of a specified organism can be compared to methylation levels of the same regions of the genome in other organisms. This comparison may allow differentiation between cancer types, or simply provide an additional filter to confirm the presence of cancer (eg, to identify false positives). Accordingly, based on said comparison, it can be determined whether the body has a first type cancer, no cancer, or a second type cancer.

Other organisms (along with the one being tested) can be clustered using the region's methylation levels. Accordingly, comparison of region methylation levels can be used to determine which cluster an organism belongs to. Clustering can also use normalized CNA values for regions for which the presence of a deletion or amplification has been determined, as described above. Also, when splitting into clusters, the corresponding methylation densities of hypermethylated CpG islands can be used.

To illustrate the principle on which this method is based, the authors of the present invention give an example of the application of logistic regression to classify two unknown samples. The purpose of this classification was to determine whether these two samples belong to HCC or non-HCC cancer. A training set of samples was constructed that included 23 plasma samples from patients with HCC and 18 samples from patients with cancer other than HCC. Thus, the training set included 41 cases in total. In this example, 13 traits were selected, including five CpG island methylation traits (X1-X5), six methylation traits for 1 Mb regions (X6-X11), and 2 CNA traits for 1 Mb regions (X12-X13). CpG methylation features were selected based on the criterion of having a Z-score of >3 or <-3 in at least 15 cases from the training set. Signs of methylation of 1 Mb regions were selected based on the criterion for the presence of a Z-score >3 or <-3

- 66 042157 in at least 39 cases from the training set. CNA features were selected based on the criterion of having a Z-score of >3 or <-3 in at least 20 cases. Logistic regression analysis was performed on samples of the specified training set to determine the regression coefficient for each of the features (X1-X13). Features with larger regression coefficients (whether in a positive or negative sense) allow better differentiation between HCC samples and non-HCC samples. Z-scores for the respective features in each case were used as input values of the independent variables. Two plasma samples were then examined for 13 features, one from a patient with HCC (TBR36) and one from a patient with lung cancer (TBR177).

In this analysis, to classify the type of cancer, it was assumed that these two samples were obtained from patients suffering from cancers of unknown origin. For each sample, the Z-scores of the respective trait were substituted into a logistic regression equation to determine the natural logarithm of the odds ratio (In(odds ratio)), where said odds ratio represents the ratio of the probabilities of having HCC versus not having HCC (HCC/non-HCC).

In table. 7 shows the regression coefficients of the logistic regression equation for 13 features. Also shown are the Z-scores of the respective features for the two tested cases (TBR36 and TBR177). The log 1p(odds ratio) of the FCC for TBR36 and TBR177 was 37.03 and -4.7, respectively. Based on the indicated odds ratios, the probability that plasma samples were obtained from patients with HCC was >99.9% and 1%, respectively. Briefly, there was a high probability that TBR36 was a sample from an HCC patient and a low probability that TBR177 was a sample from a HCC patient.

Table 7 sign Regression coefficient Z-score of the corresponding feature TBR36 TBR177 XI -2.9575 14.8 0 X2 2.2534 21.3 0 xs -1.5099 6.1 0 X4 -0.236 34.0 0 X5 0.7426 17.3 0 X6 -0.6682 -26.3 -1.5 X7 -0.2828 -13.9 -2.6 X8 -0.7281 -9.4 -4.4 X9 1.0581 -7.8 -3.7 x10 0.3877 -20.8 -4.3 XII 0.3534 -15.5 -3.1 X12 -1.1826 4.8 3.3 X13 -0.3805 -11.7 -1.4 1p(odds ratio) 37.03 -4.37463

In other embodiments, hierarchical cluster regression, classification tree analysis, and other regression models may be used to determine the probable primary origin of a cancer.

XII. Materials and methods

A. Preparation of bisulfite-treated DNA libraries and sequencing.

Genomic DNA (5 μg) supplemented with 0.5% (w/w) unmethylated lambda DNA (Promega) was fragmented using the Covaris S220 system (Covaris) into approximately 200 bp fragments. DNA libraries were prepared using a Paired-End Sequencing Sample Preparation Kit (Illumina) according to the manufacturer's instructions, except that methylated adapters (Illumina) were ligated into the DNA fragments. After two rounds of purification using AMPure XP magnetic beads (Beckman Coulter), the ligation products were split into 2 fragments, one of which was subjected to 2 rounds of bisulfite modification using the EpiTect Bisulfite Kit (Qiagen). Unmethylated cytosines in the CpG sites of the inserts were converted to uracils, while methylated cytosines remained unchanged. DNA molecules with ligated adapters, treated or not treated with sodium bisulfite, were enriched by performing 10 cycles of PCR using the following composition: 2.5 U PfuTurboCx Hotstart DNA polymerase (Agilent Technologies), 1X PfuTurboCx reaction buffer, 25 μM dNTP, 1 μl PCR primer PE 1.0 and 1 µl of PCR primer PE 2.0 (Illumina) in a reaction volume of 50 µl. The following thermal cycling profile was used: 95°C for 2 minutes, 98°C for 30 s, then 10 cycles at 98°C for 15 s, 60°C for 30 s and 72°C for 4 minutes, s final step at 72°C for 10 minutes (R Lister, et al. 2009 Nature; 462: 315-322). PCR products were purified using AMPure XP magnetic beads.

- 67 042157

Calibration fragmented lambda DNA (25 pg per ml plasma) was added to plasma DNA extracted from 3.2-4 ml of maternal plasma samples and the library was constructed as described above (RWK Chiu et al. 2011 BMJ; 342: c7401). After ligation of the methylated adapters, the ligation products were split into 2 halves and part was processed in 2 rounds of bisulfite modification. The bisulfite-treated or raw ligation products were then enriched by performing 10 cycles of PCR as described above.

Libraries of bisulfite-treated or untreated DNA were sequenced at 75 bp. in twin-end format using HiSeq2000 instruments (Illumina).

DNA clusters were generated using the Paired-End Cluster Generation Kit v3 using the cBot tool (Illumina). Real-time image analysis and base calling were performed using HiSeq Control Software (HCS) v1.4 and Real Time Analysis (RTA) Software v1.13 (Illumina), which performed automated matrix calculation and phasing based on PhiX v3 calibration control sequenced with DNA libraries.

B. Sequence alignment and identification of methylated cytosines.

After base recognition, adapter sequences and low quality bases (i.e. quality score < 20) at the ends of the fragments were removed. Clipped FASTQ reads were then processed using a methylation data analysis system called Methy-Pipe (P Jiang, et al. Methy-Pipe: An integrated bioinformatics data analysis pipeline for whole genome methylome analysis (Methy-Pipe: an integrated bioinformatics data analysis pipeline for whole genome methylome analysis). for genome-wide analysis of methylome), presented at the International Institute of Electrical and Electronics Engineers (IEEE International Conference on Bioinformatics and Biomedicine Workshops), Hong Kong, December 18-21, 2010). To align bisulfite-converted sequencing reads, the present inventors first converted all cytosine residues to thymines in silico, on Watson-Crick strands individually, using a reference human genome (NCBI version 36/hg18). The present inventors then converted each cytosine to thymine in silico in all processed reads and retained information about the location of each converted residue. SOAP2 (R Li, et al. 2009 Bioinformatics; 25: 1966-1967) was used to align converted reads to two pre-converted human reference genomes, with a maximum of two mismatches for each read aligned. Only reads mapping with unique genomic localization were selected. Ambiguous reads that mapped to both Watson-Crick strands and duplicated (clonal) reads with the same start and end positions in the genome were removed. Sequencing reads with inserts <600 bp saved for analysis of methylation and size.

Cytosine residues in the context of CpG dinucleotides have been the main target of subsequent DNA methylation studies. After alignment, cytosines originally present on the sequencing reads were reconstructed based on the location information preserved by in silico conversion. Reduced cytosines in CpG dinucleotides were counted as methylated. Thymines in CpG dinucleotides were counted as unmethylated. The unmethylated lambda DNA included in the preparation of the library was used as an internal control to evaluate the effectiveness of the modification with sodium bisulfite. All cytosines on lambda DNA were to be converted to thymines with a bisulfite conversion efficiency of 100%.

XIII. conclusions

The use of the embodiments described herein allows screening, detection, monitoring, or prediction of a cancer in a non-invasive manner using, for example, plasma from a subject. It is also possible to perform prenatal screening, obtain a diagnosis, study or monitor the fetus by establishing the methylation profile of fetal DNA in maternal plasma. To illustrate the effectiveness of this method, the authors of the present invention have shown that information that is usually obtained by examining placental tissues can be determined directly from maternal plasma. For example, it was possible to determine the imprinting status of gene loci, identify loci with differential methylation in fetal and maternal DNA, and assess gestational variation in the methylation profile of gene loci through direct analysis of maternal plasma DNA. An important advantage of the method proposed by the inventors is that fetal bromide can be comprehensively assessed during pregnancy without prejudice to pregnancy or without the need for invasive sampling of fetal tissues. Given the known links between altered DNA methylation status and many pregnancy-related conditions, the method described in this study may be a valuable tool for investigating the pathophysiology and identifying biomarkers of such conditions. By focusing on imprinting loci, the present inventors have shown that both paternal and maternal fetal methylation profiles can be assessed from maternal plasma. This method may potentially be suitable for the study of imprinting diseases. Implementations can also be applied directly to

- 68 042157 prenatal assessment of fetal diseases or pregnancy-related diseases.

The present inventors have also demonstrated that whole genome bisulfite sequencing can be used to study the DNA methylation profile of placental tissues. The human genome contains approximately 28M CpG sites (C Clark et al. 2012 PLoS One; 7: e50233). The inventors' bisulfite sequencing data for CVS and a mature placenta tissue sample captured more than 80% of CpG, which is essentially a broader coverage than is provided using other high throughput platforms. For example, the Illumina Infinium HumanMethylation 27K BeadChip Array used in a previous placental tissue study (T Chu et al. 2011 PLoS One; 6: el4723) captured only 0.1% of CpG in the genome. The recently released Illumina Infinium HumanMethylation 450K BeadChip kit captured only 1.7% CpG (C Clark et al. 2012 PLoS One; 7: e50233). Because the MPS method has no limitations in terms of probe design, hybridization efficiency, or antibody capture efficiency, CpG within and outside of CpG islands can be assessed in the context of most sequences.

XIV. computer system

In any of the computer systems mentioned in this invention, any suitable number of subsystems may be used. Examples of such subsystems are shown in Fig. 33 in computing device 3300. In some embodiments, the computer system includes a separate computing device, where subsystems may be components of said computing device. In other embodiments, the computer system may include multiple computing devices, each of which is a subsystem, with internal components.

The subsystems shown in Fig. 33 are interconnected via a system bus 3375. Additional subsystems are shown, such as a 3374 printer, a 3378 keyboard, a 3379 storage device(s), a 3376 monitor that is connected to a 3382 video adapter, and others. Peripherals and input/output (I/O) devices that are connected to the 3371 I/O controller can be connected to the computer system using any number of methods known in the art, such as the 3377 serial I/O port. The 3377 serial I/O interface or the 3381 external interface (eg, Ethernet, Wi-Fi, etc.) can be used to connect the 3300 computer system to a global computer network such as the Internet, to a mouse input device, or to a scanner. Communication via the system bus 3375 allows the central processing unit 3373 to communicate with each subsystem and control the execution of instructions from the system memory 3372 or storage device(s) 3379 (eg, a hard drive), as well as to exchange information between subsystems. 3372 system memory and/or storage device(s)

3379 may include a computer readable medium. Any of the values mentioned in this document may be passed from one component to another component and may be inferred to be presented to the user.

The computer system may include many of the same components or subsystems, for example, connected to each other through an external interface 3381 or through an internal interface. In some embodiments, computer systems, subsystems, or devices may communicate over a network. In such cases, one computer may be considered a client and the other computer a server, and each of these computers may be part of the same computer system. Both the client and the server may include multiple systems, subsystems, or components.

It should be understood that any of the embodiments of the present invention can be implemented in the form of control device logic circuits using hardware (for example, an application-specific integrated circuit or a field programmable gate array) and / or using computer software, with a broadly programmable processor in modular or integral form. As used herein, a processor includes a multi-core processor on the same integrated chip, or multiple processing modules located on a single printed circuit board or networked together. Based on the present description and the teachings herein, those of ordinary skill in the art will appreciate and understand other ways and/or methods for implementing embodiments of the present invention using hardware and hardware/software combinations.

Any software components or functions described in the present invention may be implemented as program code to be executed by a processor using any suitable programming language such as, for example, Java, C++, or Perl, using, for example, conventional or object-based oriented methods. The program code may be stored as a series of instructions or instructions on a computer readable medium for storage and/or transmission, suitable media include random access memory (RAM), read only memory (ROM), magnetic media such as a hard drive or floppy disk, or optical media. , such

- 69 042157 as compact disc (CD) or DVD (digital universal disc), flash memory, etc. Computer-readable media can be any combination of such storage or transmission devices.

Such programs may also be encoded and transmitted using carrier signals designed for transmission over wired, optical, and/or wireless networks in accordance with a variety of protocols, including the Internet. Thus, computer-readable media according to an embodiment of the present invention can be obtained by using an information signal encoded by such programs. Computer-readable media encoded with program code may be packaged with a compatible device or provided separately from other devices (eg, via Internet download). Any such computer-readable medium may be stored on/in a single computer program product (e.g., on a hard disk drive, CD-ROM, or as part of an entire computer system) and may be present in different software products or on different software products within a system or network. The computer system may include a monitor, printer, or other suitable display device to provide the user with any of the results mentioned in this invention.

Any of the methods described in this invention can be fully or partially implemented using a computer system that includes one or more processors that can be configured to perform these steps. Thus, embodiments may be directed to computer systems configured to perform the steps of any of the methods described herein, possibly with different components performing the respective steps or the respective group of steps. Although the steps according to the methods described herein have been presented in numbered form, these steps may be performed simultaneously or in a different order. In addition, step data blocks can be used with other step blocks from other methods. Also, the entire step or blocks of the step may be optional. In addition, any of the steps of any of the methods can be implemented using modules, circuits, or other means to implement these steps.

The specific details of particular embodiments may be combined in any suitable manner without departing from the spirit and scope of the embodiments of the present invention. However, other embodiments of the present invention may be directed to specific implementation options relating to each individual aspect, or to specific combinations of these individual aspects.

The previous description of exemplary embodiments of the present invention has been presented for purposes of illustration and description. It is not intended to be exhaustive or to limit the present invention to the exact form described; many modifications and variations are possible in light of the ideas described above. Embodiments have been selected and described in order to best explain the principles of the present invention and its practical application, in order to enable others skilled in the art to best apply the present invention in various implementations and with various modifications suitable for a particular intended application.

The use of the singular form, including in combination with the definition specified(specified/specified), is intended to mean one or more, unless otherwise expressly stated.

All patents, patent applications, publications, and patent disclosures mentioned herein are incorporated by reference in their entirety in all respects. None of these sources are considered prior art.

- 70 042157

Table S2A

List of 100 most hypermethylated regions identified in first trimester chorionic villus sample and maternal blood cells

Chromosome Start End Size (p.o.) Maternal blood cells CVS Meanings Methylation difference chrl3 113063600 113064100 500 0.009 0.9 3.67E-15 0.891 chr6 36279700 36280200 500 0.0068 0.8957 2.39E-22 0.8889 chrl6 66876000 66876500 500 0.0327 0.9211 3.82E-21 0.8884 chrlO 163500 164000 500 0.0195 0.9034 3.60E-35 0.8839 chr9 3518300 3518800 500 0.0263 0.9045 1.32E-26 0.8782 chrl2 31877100 31877600 500 0.007 0.8784 3.08E-22 0.8714 chr22 37477400 37478400 1000 0.0152 0.8848 0.00Е+00 0.8696 chr4 148940500 148941000 500 0.0055 0.8717 4.40E-29 0.8662 chr5 131836300 131836800 500 0.075 0.9403 1.54Е-10 0.8653 chrl7 26661700 26663600 1900 0.0187 0.875 2.95E-38 0.8563 chr2 105758600 105759600 1000 0.0305 0.8828 1D9E-53 0.8523 chr22 39188800 39189800 1000 0 0.8514 2.05Е-46 0.8514 chr3 153443900 153444900 1000 0.0436 0.8945 5.43E-34 0.8509 chr6 25149600 25150600 1000 0.0135 0.8632 0.00Е+00 0.8497 chr5 98296800 98297300 500 0.0432 0.8925 4.97E-23 0.8493 chr7 150679900 150680400 500 0.0496 0.8944 6.50E-17 0.8448 chr7 107563100 107563600 500 0.0495 0.8895 9.58E-26 0.84 chr7 37348300 37349300 1000 0.0012 0.8409 0.00Е+00 0.8397 chrl4 58837800 58838300 500 0.0097 0.848 3.35E-16 0.8383 chr6 119238100 119238600 500 0.0899 0.928 2.38E-19 0.8381 chrl5 93669900 93670400 500 0.0753 0.913 2.19E-10 0.8377 chrl7 26669200 26670200 1000 0.0221 0.859 1.44Е-29 0.8369 chr2 88108100 88108600 500 0.075 0.9109 3.55E-17 0.8359 chrl3 98363800 98364300 500 0.11 0.9457 1.28E-11 0.8357 chr!6 66948000 66948500 500 0.0331 0.8685 0.00Е+00 0.8354

- 71 042157

chr6 42098000 42098500 500 0.0484 0.8835 3.73E-16 0.8351 chr3 129876000 129876500 500 0.0565 0.8897 8.81E-17 0.8332 chr3 142700300 142700800 500 0.0063 0.8393 2.59E-22 0.833 chr8 145883800 145884300 500 0.0392 0.872 0.00E+00 0.8328 chrlO 8320700 8321200 500 0.0566 0.8871 9.40E-09 0.8305 chr3 120438100 120438600 500 0.102 0.9292 7.09E-16 0.8272 chr3 173792600 173793100 500 0.0182 0.8453 2.84E-39 0.8271 chrl7 40320700 40321200 500 0.0539 0.8788 6.50E-30 0.8249 chrl 5 72076200 72076700 500 0.0299 0.8525 6.48E-10 0.8226 chrl6 29663900 29665400 1500 0.0081 0.8305 0.00E+00 0.8224 chrll 66961100 66962100 1000 0.0489 0.8712 0.00E+00 0.8223 chr9 27083100 27084100 1000 0.097 0.9177 2.37E-55 0.8207 chr9 111249600 111250100 500 0.0613 0.8795 l,99E-20 0.8182 chrl4 101412400 101412900 500 0 0.8167 8.26E-32 0.8167 chrl 242549200 242549700 500 0 0.8155 3.50E-21 0.8155 chr8 38642800 38643300 500 0.0191 0.8346 3.22E-41 0.8155 chr4 85893600 85894100 500 0.0394 0.8533 1.45E-15 0.8139 chr5 142368600 142369100 500 0.0385 0.8523 1D8E-18 0.8138 chr8 130969500 130970000 500 0.069 0.8824 2.42E-24 0.8134 chr2 196783900 196784400 500 0 0.8123 3.63E-40 0.8123 chrl 6 49258100 49258600 500 0.0733 0.8851 4.29E-18 0.8118

chrl 232601200 232601700 500 0.0594 0.8707 1.73E-13 0.8113 chrl 109039500 109040000 500 0.0366 0.8471 l,07E-ll 0.8105 chrl 7 59491300 59491800 500 0.0662 0.8758 2D5E-17 0.8096 chr21 42194100 42194600 500 0.11 0.9182 1.61E-12 0.8082 chr9 116174500 116175500 1000 0.0062 0.8132 l,98E-60 0.807 chrl 5 73429200 73429700 500 0 0.8066 9.81E-33 0.8066 chr6 157462800 157463300 500 0.0758 0.8819 7.94E-16 0.8061 chr3 16858500 16859500 1000 0.0021 0.8068 4.76E-68 0.8047 chr9 96662800 96663300 500 0.0614 0.8651 6.79E-28 0.8037 chr9 88143000 88143500 500 0.1538 0.9559 7.43E-09 0.8021 chrl 9 16090000 16091000 1000 0.0899 0.8904 l,60E-53 0.8005 chrl 5 29436300 29437300 1000 0.0553 0.8556 1D8E-80 0.8003 chrll 77816100 77816600 500 0.1069 0.9068 2.31E-17 0.7999 chrlO 30346800 30347300 500 0.1212 0.9211 7.48E-07 0.7999 chrl 89510300 89511300 1000 0.0203 0.8191 3.53E-77 0.7988 chr3 125986100 125986600 500 0.1686 0.9674 5.24E-22 0.7988 chrl9 60162800 60163300 500 0.0127 0.8113 9.99E-19 0.7986 chrl 6 73655900 73656900 1000 0.082 0.8806 4.48E-41 0.7986 chrl 6 30104300 30105800 1500 0.0298 0.8282 0.00E+00 0.7984 chrlO 118642400 118642900 500 0.0588 0.8571 8.63E-11 0.7983 chrl 6 4495000 4496000 1000 0.0632 0.8615 2.27E-44 0.7983 chrl 2048300 2048800 500 0.0309 0.8289 1D9E-80 0.798 chr2 136481800 136482800 1000 0.0554 0.8533 8.50E-48 0.7979 chrlO 29959200 '29959700 500 0.1429 0.94 2.60E-08 0.7971 chr6 139642400 139642900 500 0.0618 0.8585 2D6E-29 0.7967 chrl4 69825300 69825800 500 0.0654 0.8615 6.85E-14 0.7961 chr8 49739700 49740200 500 0.0324 0.828 2.88E-30 0.7956 chrl 7 42205700 42206200 500 0.057 0.852 2D1E-30 0.795 chr4 77445300 77445800 500 0.0442 0.8377 l,79E-35 0.7935

- 72 042157

chrl 7 53762700 53766300 3600 0.0003 0.7926 0.00E+00 0.7923 chrl 7 44269900 44270400 500 0.026 0.8182 3.49E-21 0.7922 chr6 42462700 42463200 500 0.0761 0.8678 4.74E-22 0.7917 chr2 23396200 23396700 500 0.0333 0.8235 1.25E-14 0.7902 chr9 100921100 100921600 500 0.0244 0.814 3.32E-21 0.7896 chr7 74016100 74016600 500 0.1442 0.9333 6.74E-10 0.7891 chr6 157879000 157879500 500 0.133 0.9219 6.36E-17 0.7889 chr3 3189400 3190400 1000 0.0693 0.8571 l,38E-24 0.7878 chrl6 29581500 29584500 3000 0.0081 0.7956 0.00E+00 0.7875 chrl 7 42201800 42202800 1000 0.0884 0.8751 0.00E+00 0.7867 chrll 94257000 94257500 500 0.1122 0.8986 4.29E-10 0.7864 chrlO 14741600 14742100 500 0.0139 0.8 l,73E-20 0.7861 chr21 33826900 33827400 500 0.0879 0.8739 2.81E-11 0.786 chr4 130057200 130057700 500 0.0893 0.875 1.76E-13 0.7857 chr21 35343400 35343900 500 0 0.7853 7.43E-18 0.7853 chrl2 105372800 105373300 500 0.0923 0.8767 8.67E-22 0.7844 chr5 10799800 10800300 500 0.1429 0.9263 8.2IE-17 0.7834 chr5 16753100 16753600 500 0.041 0.8241 l,40E-15 0.7831 chr3 135746000 135746500 500 0.1429 0.9259 2.86E-09 0.783 chr6 53708300 53708800 500 0.0412 0.8235 2.74E-31 0.7823 chr2 128122900 128123400 500 0.0634 0.8455 4.82E-21 0.7821 chr5 150574200 150574700 500 0.0876 0.8696 l,56E-20 0.782

chrl6 84326000 84327000 1000 0.1071 0.8891 3.58E-61 0.782 chrl 26744500 26745500 1000 0.0336 0.8152 0.00E+00 0.7816 chr2 234882000 234882500 500 0.0392 0.819 7.63E-14 0.7798

Table S2B

List of the 100 Most Hypomethylated Regions Identified in Chorionic Villus Sample and Maternal Blood Cells in the First Trimester

Chromosome Start End Size (p.o.) Maternal blood cells CVS Meanings Methylation difference chrl 8 12217500 12218500 1000 0.9873 0 3.05Е-25 0.9873 chrl 7 22885400 22885900 500 0.9714 0.0161 8.92E-12 0.9553 chr3 184827100 184827600 500 0.9875 0.033 4.79E-16 0.9545 chr5 148968300 148968800 500 0.98 0.0426 6.70E-09 0.9374 chrlO 104794500 104795000 500 0.973 0.0385 9,ZZE-10 0.9345 chr4 84977900 84978400 500 0.9643 0.0417 2.98Е-08 0.9226 chr3 180395300 180395800 500 0.9877 0.0667 6.72Е-08 0.921 chr2 138908300 138908800 500 0.939 0.0208 Ι,ΙΟΕ-16 0.9182 chr6 139873100 139873600 500 0.9667 0.0526 1.29E-07 0.914 chr8 59604700 59605200 500 0.9468 0.033 2.88Е-14 0.9138 chr6 167622300 167622800 500 0.9452 0.0316 3.86E-14 0.9136. chr3 175701300 175701800 500 0.9846 0.0735 7.43E-10 0.9111 chrl3 59246400 59246900 500 0.9402 0.0313 2.31E-11 0.9089 chrl2 71263600 71264100 500 0.9296 0.0213 1.08E-08 0.9083 chr5 39459400 39459900 500 0.9219 0.014 5.01Е-22 0.9079 chrl 7 24904700 24905200 500 0.9161 0.0092 5.04Е-35 0.9069 chrl 2 31889900 31890400 500 0.9524 0.0465 6.78E-13 0.9059 chr3 152897800 152898300 500 0.9402 0.0345 1.70E-17 0.9057 chrl 40378700 40379200 500 0.9565 0.0526 3.31E-09 0.9039

- 73 042157

chrl2 43979300 43979800 500 0.952 0.05 6.68E-13 0.902 chrl 8 1395900 1397400 1500 0.9308 0.0293 0.00E+00 0.9015 chrl 223482900 223483400 500 0.9579 0.0575 3.36E-24 0.9004 chr9 130357000 130357500 500 0.9282 0.0286 9D9E-13 0.8996 chr3 72878300 72878800 500 0.9612 0.0625 8.20E-14 0.8987 chr7 84347200 84348700 1500 0.9401 0.0418 0.00E+00 0.8983 chrl 5 37317500 37318000 500 0.9358 0.0385 7.58E-14 0.8973 chr8 42528600 42529100 500 0.9302 0.0337 1.73E-14 0.8965 chr6 134914000 134914500 500 0.9037 0.0076 4.84E-21 0.8961 chrl3 56207100 56208100 1000 0.9184 0.0245 0.00E+00 0.894 chr2 209074000 209074500 500 0.9309 0.037 6DZE-27 0.8938 chrl2 74021100 74022100 1000 0.9513 0.058 0.00E+00 0.8933 chr4 118939300 118939800 500 0.9192 0.0276 6.58E-27 0.8916 chr5 12626600 12628600 2000 0.9266 0.0355 0.00E+00 0.8911 chr5 105517300 105518300 1000 0.927 0.0359 0.00E+00 0.891 chrl2 70056300 70057300 1000 0.9488 0.0609 0.00E+00 0.888 chr6 153238200 153239200 1000 0.9123 0.0244 0.00E+00 0.8879 chrl 7 60374800 60375300 500 0.9655 0.0777 3.64E-14 0.8878 chrl4 68272700 68273200 500 0.9389 0.0523 l,23E-22 0.8866 chrl 9 54533800 54534800 1000 0.9117 0.0262 0.00E+00 0.8855 chrl2 15392200 15393200 1000 0.9307 0.0457 0.00E+00 0.885 chrl 212517400 212517900 500 0.9266 0.0417 9.81E-12 0.8849 chrlO 49344400 49345400 1000 0.9422 0.0579 0.00E+00 0.8844 chr3 47410400 47410900 500 0.9213 0.0381 5.59E-16 0.8832 chr3 879500 880000 500 0.9455 0.0625 8.06E-06 0.883 chr2 31572400 31573400 1000 0.9176 0.0357 0.00E+00 0.8819

chrl 89131200 89131700 500 0.9314 0.0498 5D5E-70 0.8816 chr8 94832000 94832500 500 0.9156 0.0351 2D6E-65 0.8805 chr7 14008300 14009800 1500 0.9349 0.0545 0.00E+00 0.8804 chrl2 12971300 12972300 1000 0.9361 0.0559 0.00E+00 0.8802 chr5 43114700 43115200 500 0.9638 0.0842 1.79E-13 0.8796 chrll 107872400 107872900 500 0.9472 0.0677 2.31E-32 0.8794 chr8 49757600 49758100 500 0.9048 0.0269 ZD5E-52 0.8779 chrl3 33106400 33106900 500 0.9384 0.0606 9.54E-15 0.8778 chr3 190658800 190659300 500 0.9388 0.0617 2.71E-22 0.877 chrl 181508000 181508500 500 0.9259 0.0495 3.78E-15 0.8764 chrl 180436900 180437400 500 0.9412 0.0652 2.36E-13 0.876 chr6 122642800 122643800 1000 0.9218 0.0458 0.00E+00 0.8759 chr5 166429300 166429800 500 0.9551 0.08 5.26E-05 0.8751 chrl 2 14972900 14973400 500 0.9483 0.0733 2D0E-18 0.8749 chr5 123933900 123934400 500 0.943 0.0683 1D2E-39 0.8746 chr2 15969400 15970400 1000 0.8939 0.0196 9.43E-46 0.8743 chr3 167635200 167636200 1000 0.9363 0.0625 9.03E-41 0.8738 chr5 159442700 159443200 500 0.9174 0.044 6.27E-14 0.8734 chr4 48027200 48027700 ‘500 0.9839 0.1111 8.89E-06 0.8728 chr6 140071500 140072000 500 0.9234 0.0506 4.39E-33 0.8728 chrlO 22356300 22356800 500 0.9548 0.0822 l,04E-18 0.8726 chr8 61007300 61007800 500 0.9197 0.0476 1.24E-15 0.8721 chrll 95463500 95464000 500 0.9348 0.0629 1D6E-20 0.8718 chr2 216399800 216400300 500 0.938 0.0667 5.98E-06 0.8713

- 74 042157

chrl 8 57359700 57360200 500 0.9293 0.0584 1.77E-19 0.871 chr3 102734400 102734900 500 0.8917 0.0207 7.94E-22 0.871 chrl 173605700 173606200 500 0.96 0.0891 5.46E-13 0.8709 chr2 86993700 86995700 2000 0.8965 0.0261 0.00E+00 0.8704 chr3 162621100 162621600 500 0.9226 0.0526 7.89E-38 0.8699 chrl2 10144800 10145300 500 0.929 0.0598 3.45E-17 0.8691 chr3 113855100 113855600 500 0.9667 0.0982 3.97E-14 0.8685 chr2 156958200 156959200 1000 0.9252 0.0571 8.89E-50 0.8681 chr2 55775000 55776000 1000 0.9159 0.0483 0.00E+00 0.8676 chr6 124898400 124898900 500 0.8987 0.0313 1.91E-15 0.8675 chr5 42003700 42004700 1000 0.9262 0.0588 0.00E+00 0.8674 chr3 24162200 24162700 500 0.883 0.0161 l,75E-27 0.8668 chr6 35394000 35395000 1000 0.9204 0.0539 0.00E+00 0.8665 chrl 7 8451800 8453300 1500 0.9376 0.0714 0.00E+00 0.8662 chrl4 53487700 53488700 1000 0.9013 0.0353 0.00E+00 0.866 chr7 98572800 98573300 500 0.9651 0.0995 8.37E-26 0.8656 chr6 52298700 52299200 500 0.9427 0.0772 l,31E-28 0.8655 chr6 159047900 159048400 500 0.908 0.0426 3.34E-08 0.8655 chrl4 22152600 22153100 500 0.9085 0.0435 4.43E-17 0.865 chrl2 103285000 103285500 500 0.9321 0.0674 0.00E+00 0.8647 chr7 43302200 43302700 500 0.968 0.1037 6.40E-16 0.8643 chrl4 22247400 22247900 500 0.9804 0.1163 3.83E-06 0.8641 chr2 66780900 66781400 500 0.9355 0.0714 8.37E-09 0.8641 chrl2 97393000 97393500 500 0.9045 0.0408 3.46E-21 0.8637 chr5 162797900 162798900 1000 0.9271 0.0635 l,75E-57 0.8636 chr2 83598400 83599400 1000 0.9354 0.0719 0.00E+00 0.8635 chrll 111358800 111359300 500 0.9156 0.0523 4.15E-24 0.8632

chrll 104891100 104892600 1500 0.9164 0.0533 0.00E+00 0.863 chrl 184583600 184584100 500 0.9647 0.1026 3.02E-13 0.8621 chr5 132350500 132351500 1000 0.9042 0.0426 l,86E-36 0.8616 chr5 53268300 53268800 500 0.972 0.1111 4.76E-16 0.8609

Table S2C

List of the 100 Most Hypermethylated Regions Identified in the Third Trimester in Placental Tissue and in Maternal Blood Cells

Chromosome Start End Size (p.o.) To mature placenta Meanings Methylation difference chr4 78129700 78130200 500 0.0488 0.9747 3.97E-33 0.926 chr5 131467400 131467900 500 0.0213 0.9275 7D0E-27 0.9063 chrl 7 26661700 26663600 1900 0.0187 0.9226 1.79E-41 0.9039 chr4 148940500 148941000 500 0.0055 0.9079 1.82E-29 0.9024 chr9 100921100 100921600 500 0.0244 0.9242 1.38E-25 0.8998 chr6 137114200 137114700 500 0 0.8934 8.87E-14 0.8934 chr3 173792600 173793100 500 0.0182 0.9091 1.70E-42 0.8908 chr5 98296800 98297300 500 0.0432 0.9333 2.58E-23 0.8901 chrl 2 44898000 44898500 500 0 0.8889 4.47E-11 0.8889 chr3 197328900 197329400 500 0.0169 0.9048 5.55Е-10 0.8878 chr8 49739700 49740200 500 0.0324 0.9194 5.71E-34 0.887 chrl 2 122279300 122279800 500 0.0135 0.8969 3.46E-21 0.8834 chrl 7 43092200 43092700 500 0 0.8824 4.34Е-10 0.8824 chr7 107563100 107563600 500 0.0495 0.931 1.05Е-28 0.8815

- 75 042157

chrll 72543200 72543700 500 0.0377 0.9167 2.94E-09 0.8789 chrl4 58837800 58838300 500 0.0097 0.886 9D6E-18 0.8763 chr3 153443900 153444900 1000 0.0436 0.9197 6.24E-39 0.876 chr3 16953200 16953700 500 0.0896 0.9655 6.78E-09 0.876 chrl 7 42205700 42206200 500 0.057 0.933 1DZE-31 0.8759 chr6 53217600 53218100 500 0.0818 0.9571 1.54E-19 0.8754 chr3 112749000 112749500 500 0.0403 0.9154 4D1E-22 0.8752 chr8 22453700 22454200 500 0.003 0.8765 l,64E-50 0.8735 chrl 162860900 162861400 500 0.023 0.8932 8.37E-14 0.8702 chr6 36279700 36280200 500 0.0068 0.8762 1D4E-21 0.8694 chr5 80962500 80963000 500 0 0.8679 2.08E-15 0.8679 chrl 6 11312500 11313000 500 0 0.8679 2D4E-10 0.8679 chrl 6 29663900 29665400 1500 0.0081 0.8759 0.00E+00 0.8678 chr3 120438100 120438600 500 0.102 0.9639 2.98E-15 0.8618 chr8 134157000 134157500 500 0.0625 0.9219 6D0E-20 0.8594 chr6 42620900 42621400 500 0 0.8571 5.68E-08 0.8571 chr5 131836300 131836800 500 0.075 0.9315 l,26E-10 0.8565 chrl4 60290000 60290500 500 0 0.8544 5.63E-14 0.8544 chr6 42850300 42851300 1000 0.0676 0.9211 2.38E-24 0.8534 chr8 28974100 28974600 500 0.0394 0.8927 2.03E-51 0.8533 chr22 22368500 22369000 500 0.0248 0.8778 1D8E-70 0.8529 chrl4 69825300 69825800 500 0.0654 0.9174 2.73E-14 0.852 chr3 142700300 142700800 500 0.0063 0.8582 2.56E-23 0.8519 chrl 7 59491300 59491800 500 0.0662 0.9175 1.81E-16 0.8513 chrl 5 30881700 30882200 500 0.0493 0.8995 2.38E-26 0.8502 chrl 5 91496300 91496800 500 0 0.85 ZDZE-17 0.85 chrl 7 18745300 18745800 500 0.0294 0.8775 3.47E-51 0.848 chrl 5 29436500 29437000 500 0.0336 0.8811 2.62E-66 0.8476 chr2 217795300 217795800 500 0 0.8472 l,78E-22 0.8472 chrll 16328100 16328600 500 0.0278 0.875 3.43E-11 0.8472 chrl3 113063500 113064000 500 0.0102 0.8571 1.82E-15 0.8469

chr5 40472400 40472900 500 0.0197 0.8664 7.54E-35 0.8467 chrl 242549200 242549700 500 0 0.8462 8.53E-23 0.8462 chrll 58099100 58099600 500 0.0162 0.8612 4.45E-35 0.845 chr9 16020400 16020900 500 0.0132 0.8555 8.05E-23 0.8423 chr8 37550700 37551200 500 0.0093 0.8512 1D1E-16 0.8419 chr5 75722400 75722900 500 0D215 0.9627 5.97E-23 0.8411 chrl 9 60454700 60455200 500 0.0316 0.8722 2.44E-62 0.8405 chr4 99587100 99587600 500 0.0128 0.8526 1.49E-12 0.8398 chr6 25149600 25150600 1000 0.0135 0.8514 0.00E+00 0.8379 chrl 32065200 32065700 500 0 0.8371 l,09E-44 0.8371 chr7 5337200 5337700 500 0.0727 0.9098 2D8E-14 0.8371 chrl7 44269900 44270400 500 0.026 0.8621 3.94E-22 0.8361 chrl 36180800 36181300 500 0.0714 0.9067 l,23E-09 0.8352 chrl 8 10472700 10473700 1000 0.0713 0.9064 9.35E-70 0.8351 chr5 350000 350500 500 0.0297 0.8643 1.49E-16 0.8346 chr2 136481800 136482800 1000 0.0554 0.8887 l,87E-52 0.8332 chr4 89241100 89241600 500 0.1091 0.9423 l,05E-12 0.8332 chrl 7 40320700 40321200 500 0.0539 0.8859 6.94E-31 0.832 chr7 133897200 133897700 500 0.0769 0.9077 l,64E-24 0.8308 chr8 98060600 98061100 500 0.0741 0.9048 ZD1E-07 0.8307

- 76 042157

chr8 134141500 134142000 500 0 0.829 2.77E-58 0.829 chrl4 80250600 80251100 500 0.0839 0.9122 2.05E-18 0.8283 chr2 100730900 100731400 500 0.0787 0.9067 4.85E-11 0.828 chr2 88108100 88108600 500 0.075 0.901 2.10E-16 0.826 chrl9 16338500 16339500 1000 0.0011 0.8259 0.00E+00 0.8247 chr5 141791900 141792900 1000 0.0225 0.8467 0.00E+00 0.8243 chrll 116227400 116227900 500 0 0.8242 l,01E-17 0.8242 chr22 48705500 48706000 500 0.0649 0.8891 4,OOE-76 0.8242 chr9 3518300 3518800 500 0.0263 0.8493 6.66E-25 0.823 chrll 16791000 16791500 500 0.1095 0.9322 l,05E-22 0.8228 chr3 135746000 135746500 500 0.1429 0.9651 7.63E-10 0.8223 chrl 19323400 19323900 500 0.0411 0.8624 2.36E-20 0.8213 chr9 96662800 96663300 500 0.0614 0.8826 l,07E-28 0.8212 chr7 37348300 37349300 1000 0.0012 0.821 0.00E+00 0.8198 chr2 234882000 234882500 500 0.0392 0.8591 5.36E-15 0.8198 chr6 44694000 44694500 500 0.1024 0.9222 5.68E-19 0.8198 chrl 7 18320500 18321000 500 0 0.8197 2.78E-39 0.8197 chr22 28992000 28994000 2000 0.0012 0.8195 0.00E+00 0.8183 chrl 7 53762700 53766300 3600 0.0003 0.8179 0.00E+00 0.8176 chrl 114215500 114216000 500 0 0.8169 2.24E-20 0.8169 chr6 13381700 13382700 1000 0.0037 0.8206 l,03E-40 0.8169 chr5 17045400 17045900 500 0.0235 0.84 3.24E-13 0.8165 chrl2 „ 110924300 110924800 500 0.0855 0.9016 l,01E-18 0.816 chrl 200499800 200500300 500 0.011 0.8269 9.73E-24 0.8159 chr4 8311000 8311500 500 0.053 0.8687 7.82E-18 0.8157 chr8 6535300 6535800 500 0.0667 0.8824 5.25E-09 0.8157 chr6 42462700 42463200 500 0.0761 0.8919 5.78E-23 0.8157 chrl 91969900 91970400 500 0.0172 0.8325 5.23E-18 0.8152 chr2 105758600 105759600 1000 0.0305 0.8455 5D5E-52 0.815 chr21 37538500 37539000 500 0.1595 0.9745 8.81E-16 0.815 chr9 92953000 92953500 500 0.0189 0.8333 1.88E-15 0.8145

chrl 6 30104400 30105900 1500 0.0505 0.8636 0.00E+00 0.8131 chrl 234184400 234185400 1000 0.0346 0.8477 9.66E-31 0.813 chr8 19116400 19116900 500 0 0.8125 9.33E-11 0.8125 chr4 141194300 141195300 1000 0.0865 0.899 5.46E-30 0.8125

Table S2D

List of the 100 Most Hypomethylated Regions Identified in the Third Trimester in Placental Tissue and in Maternal Blood Cells

Chromosome Start End Size (p.o.) Maternal blood cells mature placenta Meanings Methylation difference chr9 40380300 40380800 500 0.9667 0 1.13E-06 0.9667 chrl 31769200 31769700 500 0.9548 0.0256 5.57E-25 0.9291 chrl 8 12217600 12218100 500 0.9873 0.0602 1.63E-19 0.9271 chr20 19704400 19704900 500 0.9426 0.018 4.34Е-18 0.9246 chrl 5 37317500 37318000 500 0.9358 0.0132 1.90E-25 0.9226 chrX 83368400 83368900 500 0.913 0 3.15E-07 0.913 chrll 27549100 27549600 500 0.9224 0.0123 3.92E-24 0.9101 chrl 8 58141500 58142000 500 0.9737 0.0645 1.07Е-09 0.9092 chrl 159897000 159897500 500 0.9067 0 2.53E-16 0.9067

- 77 042157

chr7 84347200 84348700 1500 0.9401 0.0407 0.00E+00 0.8994 chr2 216916100 216916600 500 0.9695 0.0714 2.13E-16 0.8981 chr7 144200000 144200500 500 0.9294 0.0317 l,24E-10 0.8977 chrl 241331600 241332100 500 0.9198 0.0227 0.00E+00 0.8971 chr7 123190000 123191000 1000 0.9341 0.0384 0.00E+00 0.8957 chr5 12626600 12628600 2000 0.9266 0.0321 0.00E+00 0.8944 chrl2 12971300 12972300 1000 0.9361 0.0438 0.00E+00 0.8923 chr22 20936500 20937000 500 0.9528 0.0606 l,87E-06 0.8922 chrl3 31321900 31322400 500 0.9231 0.0313 l,43E-06 0.8918 chr22 21701500 21702000 500 0.9579 0.0667 l,30E-09 0.8912 chrlO 104794400 104794900 500 1 0.1111 6.10E-09 0.8889 chr7 21835800 21836300 500 0.9156 0.0267 3.85E-13 0.8889 chrlO 16134800 16135300 500 0.95 0.0635 4.79E-10 0.8865 chr3 47410400 47410900 500 0.9213 0.0357 6.63E-17 0.8855 chrlO 49344400 49345400 1000 0.9422 0.0571 0.00E+00 0.8851 chr2 209073900 209074400 500 0.9196 0.0353 l,63E-22 0.8843 chrl 89131200 89131700 500 0.9314 0.0472 l,05E-75 0.8842 chr3 167118500 167119500 1000 0.9365 0.0527 0.00E+00 0.8838 chrl 8 1395900 1397400 1500 0.9308 0.0472 0.00E+00 0.8836 chr2 59670300 59670800 500 0.9433 0.0599 5.09E-23 0.8834 chrl4 28368900 28369400 500 0.9446 0.0619 5.03E-64 0.8827 chr3 126028800 126029300 500 0.9379 0.0556 l,83E-20 0.8823 chr9 69378900 69379900 1000 0.8816 0 6.02E-51 0.8816 chr5 105517300 105518300 1000 0.927 0.0461 0.00E+00 0.8808 chr2 31572400 31573400 1000 0.9176 0.037 0.00E+00 0.8806 chr5 42003700 42004700 1000 0.9262 0.0462 0.00E+00 0.88 chrl4 94718300 94718800 500 0.9548 0.0764 6.67E-19 0.8784 chrl 9 56417800 56418300 500 0.9615 0.0833 l,71E-06 0.8782 chr2 70183000 70183500 500 0.9694 0.0914 9.49E-39 0.878 chr4 118939300 118939800 500 0.9192 0.0412 2.20E-34 0.878 chrl3 59246400 59246900 500 0.9402 0.0633 5.40E-12 0.8769 chrl2 74021100 74022100 1000 0.9513 0.0752 0.00E+00 0.8761 chr2 173432500 173433000 500 0.9529 0.0778 5.39E-12 0.8752 chrl 6 24004400 24004900 500 0.9239 0.0488 3.25E-23 0.8751 chrl3 27596300 27597300 1000 0.9538 0.0795 0.00E+00 0.8743 chrl 5 88904300 88904800 500 0.9212 0.0481 7.69E-27 0.8731

chrl 8 12720200 12721200 1000 0.9346 0.0618 0.00E+00 0.8728 chrl 5 60975900 60976900 1000 0.9311 0.0587 0.00E+00 0.8724 chr21 39630100 39631100 1000 0.9423 0.07 4.68E-43 0.8723 chr5 123933900 123934400 500 0.943 0.0707 2.60E-38 0.8722 chr8 77382600 77383600 1000 0.9117 0.0395 0.00E+00 0.8722 chr21 32238800 32239300 500 0.9396 0.0677 1.28E-18 0.8719 chr5 175019600 175020100 500 0.9542 0.0828 4.34E-20 0.8714 chr8 134437400 134438400 1000 0.9083 0.037 4.79E-29 0.8713 chr5 69668800 69669300 500 0.9194 0.0492 2.88E-09 0.8702 chrl 60877900 60878900 1000 0.9378 0.068 0.00E+00 0.8698 chrl 6 80650400 80650900 500 0.9309 0.0611 2.49E-32 0.8698 chrl 8 59388800 59389300 500 0.9706 0.1008 4.84E-15 0.8697 chr2 15969400 15970400 1000 0.8939 0.0244 3.07E-52 0.8695 chrl3 56207100 56208100 1000 0.9184 0.0505 0.00E+00 0.868

- 78 042157

chr3 180395300 180395800 500 0.9877 0.12 l,73E-09 0.8677 chr6 153238200 153239200 1000 0.9123 0.0452 0.00E+00 0.8671 chrl8 61635100 61635600 500 0.9268 0.06 2.77E-13 0.8668 chr3 177562200 177563200 1000 0.9121 0.0455 0.00E+00 0.8666 chr4 160368300 160370200 1900 0.9272 0.0606 0.00E+00 0.8665 chr6 144626900 144627400 500 0.9114 0.046 2.10E-12 0.8654 chrl6 59885500 59886500 1000 0.9407 0.0757 l,12E-62 0.865 chrl 55667100 55667600 500 0.9095 0.0446 8.62E-39 0.8649 chr2 83598300 83599300 1000 0.9366 0.0718 0.00E+00 0.8648 chr4 105135200 105136200 1000 0.913 0.0486 0.00E+00 0.8644 chr!4 32048400 32048900 500 0.9142 0.0499 5.43E-53 0.8643 chrl 223482700 223483700 1000 0.9636 0.0997 2.69E-34 0.864 chrl4 47487700 47488200 500 0.915 0.0514 5.45E-33 0.8636 chr3 104515000 104515500 500 1 0.1373 l,08E-06 0.8627 chr7 14008300 14009800 1500 0.9349 0.0725 0.00E+00 0.8624 chrl 243134000 243135500 1500 0.9208 0.0588 0.00E+00 0.8619 chrlO 14156400 14156900 500 0.9105 0.0489 0.00E+00 0.8616 chr2 118616200 118617200 1000 0.9178 0.0565 0.00E+00 0.8613 chrl 7 8455500 8456000 500 0.8941 0.0331 1.94E-18 0.8611 chrl2 15392200 15393200 1000 0.9307 0.0697 0.00E+00 0.861 chr8 81275900 81276900 1000 0.9291 0.0684 0.00E+00 0.8606 chrl 234269300 234269800 500 0.9471 0.087 2.25E-25 0.8602 chrl 181970300 181970800 500 0.9167 0.0566 2.15E-08 0.8601 chr2 55775000 55776000 1000 0.9159 0.0559 0.00E+00 0.8599 chr3 88338000 88339000 1000 0.8909 0.0311 0.00E+00 0.8598 chr5 140078700 140079200 500 0.8852 0.0253 0.00E+00 0.8598 chr21 16720900 16721400 500 0.9317 0.0721 1.38E-15 0.8596 chrll 104891100 104892600 1500 0.9164 0.0569 0.00E+00 0.8595 chrl 184204700 184205200 500 0.9194 0.0603 8D6E-16 0.859 chr6 160732500 160733000 500 0.9191 0.0606 l,31E-10 0.8585 chr8 37134300 37134800 500 0.9151 0.0567 l,21E-26 0.8584 chrl 8 5869800 5870300 500 0.913 0.0548 l,21E-09 0.8582 chrl 98448100 98448600 500 0.9574 0D 2.08E-05 0.8574 chr3 152897800 152898300 500 0.9402 0.0828 3.28E-18 0.8574 chrl 110304000 110304500 500 0.9783 0.121 2.60E-18 0.8572 chr2 86993600 86995600 2000 0.8965 0.0395 0.00E+00 0.857 chrl 9 15428100 15430600 2500 0.9424 0.0862 0.00E+00 0.8563

chrl3 75176800 75177800 1000 0.9258 0.0697 3.09E-47 0.8561 chrl3 24126700 24127200 500 0.9498 0.0938 1.57E-17 0.856 chrl 6 28238500 28240500 2000 0.9427 0.0868 0.00E+00 0.8559 chr2 158079500 158080500 1000 0.9199 0.0642 0.00E+00 0.8557

Table S3A

List of the first 100 loci identified as hypermethylated by bisulfite sequencing for maternal plasma in the first trimester

Chromosome Start End Maternal blood cells CVS Methylation difference chr22 39189067 39189863 0 0.8444 0.8444 chrl 7 53763065 53764027 0 0.7922 0.7922 chr7 41887694 41888212 0 0.7614 0.7614

- 79 042157

chr2 1D4E+08 l,14E+08 0 0.751 0.751 chrl2 25096242 25097206 0 0.7098 0.7098 chrl 66574104 66574793 0 0.7025 0.7025 chr6 11489985 11490755 0 0.7004 0.7004 chr6 l,07E+08 l,07E+08 0 0.6978 0.6978 chrlO 30858286 30858871 0 0.6693 0.6693 chrl7 21131574 21132167 0 0.6496 0.6496 chrl 8 13454740 13455292 0 0.5468 0.5468 chrl 6 11298755 11299326 0 0.5373 0.5373 chr2 1.75E+O8 l,75E+08 0 0.5196 0.5196 chrl9 44060511 44061036 0 0.5128 0.5128 chr6 l,08E+08 l,08E+08 0 0.5 0.5 chr3 71261611 71262501 0 0.4587 0.4587 chr9 36247847 36248885 0 0.447 0.447 chrl 9 17819240 17820082 0 0.4279 0.4279 chrl 7 53769900 53770731 0 0.4102 0.4102 chrl 1D2E+08 l,12E+08 0.0002 0.6167 0.6166 chr7 l,34E+08 l,34E+08 0.0003 0.4351 0.4348 chr3 11658550 11659929 0.0004 0.4299 0.4295 chrl 7 53764417 53765963 0.0005 0.7967 0.7961 chrlO 11246762 11249052 0.0005 0.4002 0.3997 chr22 28992647 28993434 0.0006 0.8092 0.8087 chrl 5 62460278 62461007 0.0006 0.4334 0.4328 chrl 31002038 31003474 0.0007 0.5926 0.5919 chrl 9 3129246 3132159 0.0008 0.7725 0.7717 chrl2 l,21E+08 l,21E+08 0.0008 0.7303 0.7295 chrl9 12304446 12305741 0.0009 0.6986 0.6978 chr3 67788734 67789395 0.001 0.9131 0.9121 chr9 l,32E+08 l,32E+08 0.001 0.7047 0.7037 chrl9 6723370 6724479 0.001 0.689 0.688 chr3 l,84E+08 l,84E+08 0.001 0.4384 0.4374 chr2 53848089 53849214 0.001 0.4368 0.4358 chrl 7 59450886 59452113 0.0012 0.469 0.4678 chr5 l,72E+08 l,72E+08 0.0014 0.578 0.5766 chr21 35342527 35343373 0.0014 0.5392 0.5378 chr21 45164804 45165437 0.0015 0.4251 0.4236 chrX 3742417 3744601 0.0016 0.4486 0.447 chr21 45158293 45159003 0.0017 0.7799 0.7782 chr7 39839340 39839876 0.0017 0.4074 0.4057 chr2 l,75E+08 l,75E+08 0.0018 0.4816 0.4797 chrl2 l,24E+08 l,24E+08 0.0019 0.6306 0.6287

chr3 50352688 50353823 0.002 0.624 0.622 chr9 97264382 97265523 0.0021 0.5008 0.4987 chr7 64178628 64179354 0.0021 0.4088 0.4066 chr9 94767202 94767802 0.0023 0.7568 0.7544 chr5 42986308 42988304 0.0023 0.4882 0.4859 chrl 7 63854127 63854693 0.0024 0.8266 0.8242 chrl2 l,22E+08 l,22E+08 0.0024 0.4869 0.4844 chrl7 16260170 16260909 0.0026 0.6404 0.6378 chr4 39874787 39875456 0.0027 0.7233 0.7206

- 80 042157

chrl2 6441080 6441608 0.0027 0.6228 0.6201 chrl9 45015653 45016886 0.0028 0.5444 0.5416 chr6 30757752 30758823 0.0028 0.4783 0.4755 chr6 41636176 41637112 0.0028 0.4254 0.4226 chrl2 6315199 6315765 0.0029 0.4613 0.4584 chrl4 76576283 76577070 0.0029 0.4365 0.4336 chrl6 48857790 48858300 0.0031 0.5625 0.5594 chr5 l,7E+08 1.7E+O8 0.0031 0.4752 0.4721 chrl3 26897813 26898557 0.0032 0.4354 0.4322 chrl4 52753948 52754571 0.0032 0.4221 0.4189 chrl l,66E+08 l,66E+08 0.0033 0.5579 0.5545 chrl2 56157424 56158348 0.0033 0.47 0.4667 chr22 16079971 16080532 0.0034 0.6226 0.6193 chr7 1946410 1946975 0.0036 0.6826 0.6789 chrll 258799 259749 0.0036 0.5072 0.5037 chr6 13381944 13382477 0.0037 0.5945 0.5908 chr7 l,27E+08 l,27E+08 0.0037 0.5096 0.5058 chrl3 23743886 23744467 0.0037 0.4534 0.4497 chr2 l,21E+08 l,21E+08 0.0038 0.7175 0.7137 chr21 25855853 25857105 0.0039 0.4661 0.4622 chr2 43211724 43212565 0.0039 0.4345 0.4306 chrl2 l,08E+08 l,08E+08 0.0041 0.6024 0.5983 chrl5 92928924 92929575 0.0041 0.4074 0.4033 chrl9 10731043 10731636 0.0042 0.5868 0.5826 chr6 l,45E+08 l,45E+08 0.0043 0.5783 0.574 chrl 52875323 52875907 0.0044 0.4145 0.4101 chrl4 75058186 75058956 0.0045 0.602 0.5975 chrl 2 l,21E+08 l,21E+08 0.0045 0.4821 0.4776 chrl 7 76873737 76874417 0.0046 0.6012 0.5966 chr2 2.38E+08 2.38E+08 0.0049 0.7654 0.7604 chr2 l,98E+08 l,98E+08 0.0049 0.7228 0.7179 chr6 l,47E+08 l,47E+08 0.0049 0.4967 0.4918 chr9 l,36E+08 l,36E+08 0.0049 0.4584 0.4535 chrl 67545402 67546771 0.005 0.4971 0.4921 chr6 l,58E+08 l,58E+08 0.0052 0.6145 0.6093 chr3 1.7E+O8 l,7E+08 0.0052 0.5845 0.5794 chrl 2.34E+08 2.34E+08 0.0053 0.7033 0.6979 chrlO 80715722 80716751 0.0053 0.6515 0.6462 chr4 48602901 48603736 0.0053 0.6315 0.6262 chrl9 13957965 13958580 0.0053 0.599 0.5937 chrl 90081114 90082367 0.0053 0.4574 0.4521 chr2 l,06E+08 l,06E+08 0.0054 0.8858 0.8804

chrl 6 29664213 29665369 0.0054 0.8339 0.8285 chrl l,59E+08 l,59E+08 0.0054 0.7663 0.7608 chrl3 97926489 97927025 0.0054 0.6229 0.6175 chrl 41604452 41605277 0.0054 0.6011 0.5956 chr9 l,28E+08 l,28E+08 0.0054 0.5871 0.5818

-81 042157

Table S3B

List of the first 100 loci identified as hypomethylated by bisulfite sequencing for maternal plasma in the first trimester

Chromosome Start End Maternal blood cells CVS Methylation difference chrl 235771917 235772426 0.9868 0.549 0.4377 chrl 97357972 97358622 0.9835 0.4805 0.503 chrl 4490516 4491074 0.9826 0.4793 0.5032 chr4 181124168 181124671 0.9825 0.4725 0.5099 chrl 6 71908694 71909213 0.982 0.5581 0.4239 chr3 182727915 182728477 0.981 0.3577 0.6233 chr5 115339535 115340038 0.9802 0.5455 0.4347 chr3 195855575 195856122 0.9801 0.3793 0.6008 chr6 155437621 155438161 0.9799 0.5991 0.3808 chr9 20468093 20468904 0.9798 0.4271 0.5527 chrlO 90702298 90702987 0.9787 0.3324 0.6463 chrl 170581654 170582162 0.9785 0.4817 0.4968 chr3 108816849 108817794 0.9783 0.4793 0.4989 chr20 36912749 36913319 0.9783 0.5 0.4783 chrl3 72517281 72517839 0.9782 0.4855 0.4927 chrl2 103553001 103553677' 0.9774 0.492 - 0.4854 chr22 27638905 27639408 0.9766 0.5385 0.4382 chr7 17290850 17291462 0.9763 0.59 0.3863 chr6 17227866 17228510 0.976 0.4058 0.5703 chrl 5 56998547 56999107 0.9754 0.3766 0.5988 chr7 70965945 70966842 0.9753 0.5893 0.386 chr3 32159338 32160065 0.9752 0.5379 0.4372 chrl 6 17043258 17043854 0.9752 0.5521 0.4231 chrl 6 22776223 22776850 0.9752 0.5735 0.4017 chr5 169344029 169344869 0.9751 0.4211 0.5541 chrll 34324955 34325722 0.975 0.5561 0.4189 chr8 58554745 58555376 0.9747 0.5784 0.3964 chrl 153933389 153934121 0.9746 0.463 0.5116 chrl4 88003983 88004485 0.9745 0.5379 0.4366 chr3 151738501 151739120 0.9741 0.4901 0.484 chrl4 105618699 105619606 0.974 0.3457 0.6283 chrl6 24060085 24060702 0.9738 0.3991 0.5747 chr8 68941792 68942711 0.9738 0.5449 0.429 chrl 2 53208707 53209304 0.9737 0.4847 0.489 chr7 76892564 76893249 0.9736 0.5664 0.4072 chr3 69464294 69464971 0.9736 0.5893 0.3843 chrl 9 61401137 61401745 0.9732 0.4933 0.4799 chrll 124569867 124570490 0.9732 0.5136 0.4595 chrl 8 42618440 42619096 0.9732 0.5942 0.379 chr5 169398896 169399637 0.9731 0.498 0.4751 chr5 169328124 169328983 0.9731 0.572 0.401 chr20 34679880 34680448 0.9731 0.5922 0.3809 chrl 6 9042198 9042702 0.973 0.4286 0.5444 chrlO 90205044 90205701 0.973 0.4407 0.5323 chrl3 33236454 33236997 0.973 0.5906 0.3824

chrl6 | 73284579 | 73285087 | 0.9729 | 0.5602 | 0.4127

- 82 042157

chr8 29100691 29101428 0.9728 0.505 0.4678 chr2 202383851 202384447 0.9727 0.5461 0.4267 chr3 179501620 179502300 0.9722 0.5766 0.3956 chr6 107674976 107675906 0.9719 0.4434 0.5285 chr6 107880632 107881161 0.9718 0.5623 0.4095 chrl2 56350283 56350933 0.9718 0.5909 0.3809 chrl9 40636458 40637339 0.9717 0.4941 0.4776 chr2 223472599 223473287 0.9714 0.1824 0.7891 chr22 20709067 20709787 0.9714 0.5149 0.4565 chr!9 46095583 46096190 0.9713 0.5385 0.4328 chr6 90258338 90259318 0.9712 0.3415 0.6297 chr2 54598347 54598933 0.9712 0.5894 0.3819 chr3 114810453 114811493 0.9711 0.5166 0.4545 chrl9 15851125 15851654 0.9711 0.5236 0.4476 chr8 42889138 42890084 0.9711 0.5652 0.4059 chrl8 52354390 52355064 0.971 0.598 0.373 chrl 5 38206236 38207010 0.9709 0.4186 0.5523 chr7 99700554 99701110 0.9708 0.305 0.6658 chrl2 19487336 19487855 0.9708 0.4105 0.5603 chr7 87996908 87997437 0.9708 0.5462 0.4246 chr6 63628653 63629378 0.9707 0.529 0.4417 chrl5 38209108 38209618 0.9706 0.5882 0.3824 chrl 9 6623769 6624450 0.9704 0.5179 0.4526 chr2 10794513 10795242 0.9704 0.5976 0.3728 chr2 118472785 118474454 0.9704 0.5992 0.3712 chr5 57820209 57820801 0.9701 0.5815 0.3886 chrlO 100183380 100184702 0.9701 0.5826 0.3875 chr2 8151989 8152646 0.97 0.4701 0.4999 chrlO 3938374 3938914 0.9699 0.1741 0.7958 chr9 123724524 123725439 0.9697 0.57 0.3997 chrl4 89085469 89086097 0.9696 0.3278 0.6418 chrl6 14129437 14130133 0.9695 0.5304 0.4392 chr5 60746367 60747191 0.9695 0.5571 0.4124 chrl 92002953 92003729 0.9694 0.52 0.4494 chr6 31264677 31265413 0.9693 0.5135 0.4558 chr7 99317013 99318281 0.9692 0.5117 0.4574 chr8 8808867 8809422 0.9692 0.5691 0.4002 chrl 9 20052165 20052720 0.969 0.2792 0.6898 chr8 129139026 129139573 0.969 0.3458 0.6232 chrll 122314929 122315458 0.969 0.4232 0.5458 chrl3 98377663 98378165 0.9688 0.3319 0.6369 chr9 107606194 107606872 0.9688 0.449 0.5198 chr8 56096904 56097736 0.9688 0.5267 0.4422 chr7 128093836 128094339 0.9688 0.5929 0.3758 chr2 103109370 103109916 0.9686 0.3333 0.6352 chr3 101803534 101804063 0.9686 0.5027 0.4659 chrlO 69505720 69506278 0.9684 0.2515 0.7169 chrl3 26608225 26608754 0.9683 0.3614 0.6069 chrl 90993315 90993828 0.9683 0.5519 0.4164 chr6 11361243 11361801 0.9681 0.2578 0.7103

chr21 36529300 36529981 0.968 0.1944 0.7736 chr21 37813953 37814521 0.9679 0.2175 0.7505 chr2 15226273 15227211 0.9679 0.5134 0.4545 chrl 9 4102809 4103443 0.9679 0.5646 | 0.4034

- 83 042157

Table S3C

List of the first 100 loci identified as hypermethylated by bisulfite sequencing for maternal plasma in the third trimester

Chromosome Start End Maternal blood cells mature placenta Methylation difference chrl7 53763065 53764027 0.0000 0.8680 0.8680 chr22 39189067 39189863 0.0000 0.8233 0.8233 chrlO 30858286 30858871 0.0000 0.7713 0.7713 chr7 41887694 41888212 0.0000 0.7578 0.7578 chr2 1D4E+08 1D4E+08 0.0000 0.7500 0.7500 chrl2 25096242 25097206 0.0000 0.7332 0.7332 chr6 l,07E+08 l,07E+08 0.0000 0.7229 0.7229 chrl 66574104 66574793 0.0000 0.7136 0.7136 chrl6 11298755 11299326 0.0000 0.7005 0.7005 chr6 11489985 11490755 0.0000 0.6935 0.6935 chrl 8 13454740 13455292 0.0000 0.6594 0.6594 chr6 l,08E+08 l,08E+08 0.0000 0.6231 0.6231 chrX 3627885 , 3628549 0.0000 0.6133 0.6133 chrl 2 7979754 7980413 0.0000 0.6118 0.6118 chr3 71261611 71262501 0.0000 0.5938 0.5938 chrl 7 53769900 53770731 0.0000 0.5586 0.5586 chrll 1D8E+08 1D8E+08 0.0000 0.5558 0.5558 chrl9 44060511 44061036 0.0000 0.5464 0.5464 chr2 2.38E+08 2.38E+08 0.0000 0.5330 0.5330 chrl l,91E+08 l,91E+08 0.0000 0.5294 0.5294 chrl l,44E+08 l,44E+08 0.0000 0.4857 0.4857 chr2 l,75E+08 l,75E+08 0.0000 0.4785 0.4785 chr4 15366889 15367646 0.0000 0.4729 0.4729 chr2 19537237 19537737 0.0000 0.4599 0.4599 chrl 1D5E+08 1D5E+08 0.0000 0.4351 0.4351 chrl l,54E+08 l,54E+08 0.0000 0.4299 0.4299 chrl 4 51383387 51384149 0.0000 0.4186 0.4186 chrl 1D2E+08 1D2E+08 0.0002 0.5350 0.5348 chr3 11658550 11659929 0.0004 0.5579 0.5575 chrl 7 53764417 53765963 0.0005 0.7894 0.7889 chr22 28992647 28993434 0.0006 0.8053 0.8047 chr6 27214981 27215823 0.0006 0.4593 0.4587 chrl 31002038 31003474 0.0007 0.6309 0.6302 chrl2 l,21E+08 l,21E+08 0.0008 0.7360 0.7352 chrl 9 3129246 3132159 0.0008 0.7257 0.7249 chrl9 12304446 12305741 0.0009 0.6397 0.6388 chr6 28723918 28724965 0.0009 0.4344 0.4335 chrl9 6723370 6724479 0.0010 0.7280 0.7270 chr2 53848089 53849214 0.0010 0.4060 0.4050 chr9 l,32E+08 l,32E+08 0.0010 0.7558 0.7548 chr3 67788734 67789395 0.0010 0.9219 0.9209 chrl 9 18276368 18277132 0.0011 0.5136 0.5125 chrl 7 59450886 59452113 0.0012 0.5196 0.5184 chrl7 74243852 74244670 0.0012 0.4117 0.4105

- 84 042157

chr3 1.85E+O8 1.85E+O8 0.0014 0.4961 0.4948 chr21 35342527 35343373 0.0014 0.5126 0.5112 chr5 l,72E+08 l,72E+08 0.0014 0.6531 0.6516 chr21 45164804 45165437 0.0015 0.4364 0.4349 chrX 3742417 3744601 0.0016 0.7517 0.7501 chr21 45158293 45159003 0.0017 0.8180 0.8163 chr2 l,75E+08 l,75E+08 0.0018 0.6214 0.6196 chrl2 l,24E+08 l,24E+08 0.0019 0.5906 0.5887 chr3 50352688 50353823 0.0020 0.6082 0.6062 chr9 94767202 94767802 0.0023 0.8327 0.8304 chrl7 63854127 63854693 0.0024 0.7886 0.7862 chrl2 l,22E+08 l,22E+08 0.0024 0.5021 0.4997 chrl7 16260170 16260909 0.0026 0.5780 0.5754 chrl2 6441080 6441608 0.0027 0.7471 0.7444 chr4 39874787 39875456 0.0027 0.7962 0.7935 chrl 8 50536032 50536649 0.0027 0.4920 0.4893 chr6 30757752 30758823 0.0028 0.4029 0.4001 chrl9 2571763 2572292 0.0031 0.4200 0.4169 chr5 l,7E+08 l,7E+08 0.0031 0.4218 0.4187 chrl3 26897813 26898557 0.0032 0.6485 0.6453 chrl2 56157424 56158348 0.0033 0.5541 0.5508 chrl l,66E+08 l,66E+08 0.0033 0.5147 0.5113 chr22 16079971 16080532 0.0034 0.6265 0.6231 chr6 16820551 16821134 0.0035 0.4800 0.4765 chrll 258799 259749 0.0036 0.5475 0.5439 chr7 1946410 1946975 0.0036 0.8251 0.8215 chr6 13381944 13382477 0.0037 0.8221 0.8183 chr7 l,27E+08 l,27E+08 0.0037 0.4767 0.4730 chr2 l,21E+08 l,21E+08 0.0038 0.6734 0.6697 chr2 43211724 43212565 0.0039 0.4256 0.4217 chrl 5 92928924 92929575 0.0041 0.5605 0.5564 chrl 2 l,08E+08 l,08E+08 0.0041 0.7313 0.7271 chrl 9 10731043 10731636 0.0042 0.5668 0.5626 chr6 l,45E+08 l,45E+08 0.0043 0.5910 0.5867 chrl 52875323 52875907 0.0044 0.6115 0.6071 chrl 2 l,21E+08 l,21E+08 0.0045 0.5884 0.5839 chrl4 75058186 75058956 0.0045 0.6534 0.6489 chrl 7 76873737 76874417 0.0046 0.5658 0.5612 chr6 l,47E+08 l,47E+08 0.0049 0.5826 0.5777 chr2 l,98E+08 l,98E+08 0.0049 0.7944 0.7895 chr2 2.38E+08 2.38E+08 0.0049 0.7328 0.7278 chr8 l,42E+08 l,42E+08 0.0050 0.7728 0.7679 chr3 l,7E+08 l,7E+08 0.0052 0.7227 0.7176 chr6 1.58E+08 l,58E+08 0.0052 0.6389 0.6337 chr2 2.38E+08 2.38E+08 0.0052 0.4238 0.4185 chrll 56948854 56949496 0.0053 0.4484 0.4431 chr4 48602901 48603736 0.0053 0.5920 0.5867 chr5 l,31E+08 l,31E+08 0.0053 0.4858 0.4805 chrlO 80715722 80716751 0.0053 0.5249 0.5196 chrl 9 13957965 13958580 0.0053 0.4379 0.4326 chrl 2.34E+08 2.34E+08 0.0053 0.8440 0.8387

chrl3 97926489 97927025 0.0054 0.7233 0.7179 chr9 l,28E+08 l,28E+08 0.0054 0.7312 0.7258 chr2 l,06E+08 l,06E+08 0.0054 0.8513 0.8459 chr2 96556705 96557637 0.0054 0.4095 0.4041 chrl 6 29664213 29665369 0.0054 0.8837 0.8783

- 85 042157

Table S3D

List of the first 100 loci identified as hypomethylated by bisulfite sequencing for maternal plasma in the third trimester

Chromosome Start End Maternal blood cells mature placenta Methylation difference chrlO 7548948 7549483 0.9866 0.5685 0.4181 chrl 4490516 4491074 0.9826 0.5015 0.4810 chr4 l,81E+08 l,81E+08 0.9825 0.5981 0.3843 chr3 l,83E+08 l,83E+08 0.9810 0.2925 0.6886 chr3 l,96E+08 l,96E+08 0.9801 0.4643 0.5158 chr6 l,55E+08 l,55E+08 0.9799 0.4610 0.5189 chrl l,71E+08 l,71E+08 0.9785 0.5122 0.4662 chr20 . 36912749 36913319 0.9783 0.4513 0.5269 chr22 38583100 38583616 0.9783 0.5428 0.4355 chrl 19391314 19392207 0.9778 0.5273 0.4505 chr5 l,74E+08 l,74E+08 0.9770 0.5852 0.3918 chrl 9 13678906 13679531 0.9760 0.5812 0.3949 chrl4 83650790 83651395 0.9760 0.5378 0.4382 chrl 5 56998547 56999107 0.9754 0.4691 0.5063 chrl 6 22776223 22776850 0.9752 0.5114 0.4638 chr5 l,69E+08 l,69E+08 0.9751 0.4809 0.4943 chr8 58554745 58555376 0.9747 0.5977 0.3770 chrl 4 l,06E+08 l,06E+08 0.9740 0.2069 0.7671 chr8 68941792 68942711 0.9738 0.5872 0.3866 chrl 6 24060085 24060702 0.9738 0.3470 0.6268 chrl 2 53208707 53209304 0.9737 0.5278 0.4459 chr5 l,69E+08 l,69E+08 0.9731 0.5057 0.4673 chrl 6 9042198 9042702 0.9730 0.1860 0.7869 chrlO 90205044 90205701 0.9730 0.5922 0.3808 chr3 l,89E+08 l,89E+08 0.9720 0.4949 0.4771 chr6 l,08E+08 l,08E+08 0.9719 0.5825 0.3894 chr2 2.23E+08 2.23E+08 0.9714 0.3333 0.6381 chrl 9 46095583 46096190 0.9713 0.5065 0.4648 chr8 l,41E+08 l,41E+08 0.9713 0.5753 0.3959 chr6 90258338 90259318 0.9712 0.4357 0.5355 chrl3 51403556 51404069 0.9710 0.3980 0.5731 chrl 8 66875048 66875726 0.9710 0.5259 0.4451 chr7 99700554 99701110 0.9708 0.3757 0.5951 chr7 87996908 87997437 0.9708 0.5720 0.3988 chrl 9 6623769 6624450 0.9704 0.4774 0.4930 chrl 97639047 97639749 0.9701 0.4148 0.5553 chrl 6 23892096 23892772 0.9701 0.5000 0.4701 chrlO 3938374 3938914 0.9699 0.1148 0.8551 chrl 4 89085469 89086097 0.9696 0.2964 0.6732 chr8 l,29E+08 l,29E+08 0.9690 0.3565 0.6126 chrl3 98377663 98378165 0.9688 0.3123 0.6566 chr8 56096904 56097736 0.9688 0.4562 0.5126 chr2 l,03E+08 l,03E+08 0.9686 0.3459 0.6227

- 86 042157

chrl3 26608225 26608754 0.9683 0.4562 0.5121 chr2 22738157 22738760 0.9682 0.5122 0.4560 chr6 11361243 11361801 0.9681 0.2646 0.7035 chr21 36529300 36529981 0.9680 0.1829 0.7852

chr21 37813953 37814521 0.9679 0.3061 0.6619 chr2 2.43E+08 2.43E+08 0.9679 0.5750 0.3929 chr4 12413543 12414103 0.9679 0.5944 0.3735 chr3 l,27E+08 l,27E+08 0.9677 0.4030 0.5648 chr7 33509047 33509556 0.9676 0.4627 0.5048 chrl4 59284846 59285553 0.9674 0.5254 0.4420 chrl7 42623453 42624024 0.9673 0.4318 0.5355 chrl 9 6778363 6779377 0.9671 0.4416 0.5255 chr4 41798250 41798788 0.9670 0.5000 0.4670 chr5 88054080 88054588 0.9669 0.2238 0.7431 chrl6 24109379 24110289 0.9669 0.5062 0.4607 chrlO 13847159 13847895 0.9667 0.3188 0.6479 chrlO l,27E+08 l,27E+08 0.9667 0.5423 0.4244 chrl2 1D2E+08 1D2E+08 0.9663 0.3722 0.5941 chrlO 17220886 17221845 0.9662 0.4455 0.5207 chr8 5947355 5947862 0.9662 0.5171 0.4491 chr3 73740840 73741439 0.9659 0.3657 0.6002 chrl4 57945953 57946875 0.9658 0.5357 0.4301 chrl4 50905777 50906333 0.9658 0.3008 0.6650 chrl5 90275374 90276000 0.9657 0.5409 0.4248 chr22 24717299 24718197 0.9657 0.5160 0.4497 chr7 36530128 36530987 0.9656 0.5194 0.4462 chr2 l,31E+08 l,31E+08 0.9655 0.4384 0.5271 chr4 42116988 42117788 0.9654 0.5195 0.4459 chrl 2 1D6E+08 1D6E+08 0.9653 0.5594 0.4059 chr2 7491785 7492736 0.9652 0.4556 0.5097 chrl9 6599638 6600187 0.9652 0.5488 0.4163 chr6 25326803 25327398 0.9651 0.3974 0.5677 chr4 l,7E+08 l,7E+08 0.9651 0.4933 0.4718 chr7 99875338 99876155 0.9650 0.2696 0.6953 chrl4 97144328 97145208 0.9649 0.5377 0.4272 chr3 11718596 11719163 0.9649 0.5521 0.4128 chrl4 1E+08 1E+08 0.9649 0.3794 0.5855 chr7 l,5E+08 l,5E+08 0.9648 0.3327 0.6322 chrl2 56357827 56358328 0.9648 0.4217 0.5430 chrlO 8275750 8276276 0.9647 0.3100 0.6547 chrll 16999685 17000209 0.9647 0.2765 0.6882 chr22 34419356 34419861 0.9646 0.4245 0.5401 chrl8 72453151 72453725 0.9646 0.4700 0.4946 chr5 49919879 49920699 0.9645 0.3169 0.6476 chrl 24580891 24581805 0.9643 0.3565 0.6078 chr22 18233774 18234492 0.9641 0.5205 0.4436 chrl4 45356178 45356903 0.9640 0.3934 0.5706 chr3 53007193 53008661 0.9638 0.4902 0.4737 chr4 55027912 55028539 0.9637 0.5254 0.4384 chr5 l,37E+08 l,37E+08 0.9637 0.5290 0.4347 chrl 2.23E+08 2.23E+08 0.9636 0.0997 0.8640 chr7 l,35E+08 l,35E+08 0.9636 0.2959 0.6677 chr5 80350438 80351169 0.9636 0.4969 0.4667 chrl2 31889600 31890343 0.9636 0.1745 0.7891 chrl 2 8365395 8366096 0.9636 0.5721 0.3914

chrl9 15424819 15425355 0.9635 0.2836 0.6799 chrlO 10985469 10986409 0.9635 0.4877 0.4759

Claims (58)

CLAIM 1. A method of analyzing a biological sample of an organism to determine the classification of the level of cancer, where the specified biological sample contains molecules of deoxyribonucleic acid (DNA) originating from normal cells and potentially from cells associated with cancer, and at least some of the DNA molecules in a biological sample are extracellular DNA molecules, said method comprising analyzing a plurality of DNA molecules from said biological sample, wherein the analysis of the DNA molecule - 87 042157 includes determining the localization of the specified DNA molecule in the genome of the specified organism; determining whether said DNA molecule is methylated at one or more sites, wherein said determination includes performing methylation-sensitive massive parallel sequencing; for each of the many sites: determining the appropriate number of DNA molecules that are methylated at the specified site; calculating a first methylation level based on the respective amounts of DNA molecules methylated at a plurality of sites, wherein the first methylation level may correspond to a methylation index, a methylation density, or a proportion of methylated cytosines, where the methylation index refers to the proportion of DNA molecules showing methylation at one or more sites relative to the number of DNA molecules covering one or more sites, where the methylation density of a region is the number of DNA molecules at the sites of the specified region showing methylation divided by the number of DNA molecules covering the sites of the region, and where the proportion of methylated cytosines refers to the cytosine sites for which shows methylation from among the analyzed cytosine residues in the region; generating a comparison of the first methylation level with the first limit value, where the specified first limit value represents the reference methylation level, is at the specified distance from the reference methylation level, or is set to the specified reference methylation level, where the reference methylation level is determined from one or more samples, cancer level in which it is known; and determining a first cancer level classification, which is a methylation classification, based on said comparison, where the cancer level may refer to either the presence of cancer, the stage of cancer, the size of the cancerous tumor, or the presence of metastases, total tumor burden on the body, and/or another measure the severity of the cancer. 2. The method according to claim 1, characterized in that the implementation of methylation-sensitive sequencing includes the treatment of DNA molecules with sodium bisulfite; sequencing of processed DNA molecules. 3. The method of claim 2, wherein the treatment of DNA molecules with sodium bisulfite is part of a bisulfite conversion using Tet or oxidative bisulfite sequencing to detect 5-hydroxymethylcytosine. 4. The method according to claim 1, characterized in that the relative amounts of DNA molecules are determined by aligning reads of sequences obtained by methylation-sensitive sequencing. 5. The method of claim 1, wherein determining whether said DNA molecule is methylated at one or more sites comprises the use of methylation-sensitive restriction enzyme digestion, methylation-specific PCR, methylation-dependent DNA precipitation binding methylated DNA protein/peptide or monomolecular sequencing without sodium bisulfite treatment. 6. The method of claim 1, further comprising, for each of the first plurality of regions of the genome, determining the appropriate number of DNA molecules originating from said region in the biological sample; calculating the corresponding normalized value according to the corresponding number of DNA molecules originating from the specified area; comparing the corresponding normalized value with a reference value to determine if there is a deletion or amplification in the region, where the reference value can be determined from a normal sample, a value of a different haplotype, or a value for a different region; determining a first number of regions from a first plurality of regions of the genome for which the presence of a deletion or amplification has been determined; comparing said first amount with a first threshold value to determine a second cancer level classification, which is a copy number aberration classification; and using the first classification and the second classification to determine the third cancer level classification, which is the methylation and copy number aberration classifications. 7. The method of claim 6, wherein the first cutoff is a percentage of the first plurality of regions for which the presence of a deletion or amplification has been determined. 8. The method according to claim 6, characterized in that the third classification is cancer-positive only if both the first classification and the second classification indicate cancer. 9. The method according to claim 1 or 6, characterized in that if the first classification indicates the presence of a cancer in the body, then the method further includes identifying the type of cancer - 88 042157 diseases associated with the specified organism, by comparing the first level of methylation with the corresponding value determined for other organisms, while at least two of these other organisms are identified as having different types of cancers. 10. The method according to claim 6, characterized in that if the third classification indicates a cancer in the organism, then the method further comprises identifying the type of cancer associated with said organism by comparing the first number of areas with corresponding values determined for other organisms. 11. The method according to claim 1 or 6, characterized in that the calculation of the first level of methylation includes the identification of the second set of regions of the genome; identifying one or more sites within each area of the second plurality of areas; calculating a region methylation level for each region of the second plurality of regions, wherein the first methylation level corresponds to the first region of the second plurality of regions, wherein generating a comparison of the first methylation level with the first cutoff further comprises comparing each of the region's methylation levels with a corresponding cutoff for the region, including comparing the first level of methylation with the first limit value; and where the determination of the first classification of the level of cancer based on the comparison includes determining the second number of areas of the second set of areas, which were determined to have a methylation level of the area exceeding the corresponding limit value for the area; and comparing said second number of regions with a second threshold to determine a first classification. 12. The method of claim 11, wherein said regions that have been determined to have a region methylation level in excess of the corresponding region cut-off correspond to the first set of regions, and further comprising comparing the methylation levels of the region from the first set of regions with corresponding methylation levels of the region in other organisms for the first set of regions, wherein said other organisms have at least two of the following: a first type of cancer, no cancer, and a second type of cancer; and determining whether the body has a first type cancer, does not have a cancer, or has a second type cancer based on said comparison. 13. The method of claim 12, wherein the comparison comprises clustering said other organisms based on respective methylation levels of a region from a first set of regions of other organisms, wherein two of said clusters correspond to any two of a first type of cancer, no cancer, and a second type of cancer, determining which cluster an organism belongs to. 14. The method of claim 13, wherein the clustering of said other organisms uses the methylation levels of a region of said organism. 15. The method of claim 13, wherein said clusters include a first cluster corresponding to a first type of cancer, a second cluster corresponding to a second type of cancer, and a third cluster corresponding to no cancer. 16. The method of claim 13, wherein the clustering of said other organisms is further based on respective normalized values of a second set of regions for said other organisms, wherein said second set of regions corresponds to regions for which the presence of a deletion or amplification has been determined, and where the corresponding normalized value for the area is determined by the corresponding number of DNA molecules from the specified area; wherein said method further comprises for each of the second set of regions: determining the appropriate number of DNA molecules as originating from the specified area; calculating an appropriate normalized value for a corresponding number of DNA molecules from the specified region; and comparing the corresponding normalized values of the second set of areas for the specified organism with the corresponding normalized values for other organisms, within the framework of determining which cluster the specified organism belongs to. 17. The method of claim 16, wherein the clustering of said other organisms is additionally based on the respective methylation densities of the hypermethylated CpG islands, and said method further comprises, for each of the hypermethylated CpG islands: determining the appropriate methylation density, and - 89 042157 comparison of the corresponding methylation densities of hypermethylated CpG islands in a specified organism with methylation densities in other organisms in order to determine which cluster the specified organism belongs to. 18. The method of claim 11, further comprising, for each of the second plurality of regions: calculating an appropriate difference between a region's methylation level and a corresponding region limit; and calculating the corresponding probability, which is the probability of obtaining the corresponding difference by chance according to the statistical distribution; wherein determining the second number of areas from the second plurality of areas includes calculating a cumulative score including the sum of the respective probabilities. 19. The method according to claim 18, characterized in that the calculation of the specified cumulative indicator includes calculating the logarithm of the corresponding probability of each area of the second set of areas to obtain the corresponding logarithmic result; and calculating the sum including the corresponding logarithmic results. 20. The method of claim 18, wherein the respective difference of each region of the second plurality of regions is normalized to a standard deviation associated with the respective limit value for the region. 21. The method according to claim 18, characterized in that the second threshold value corresponds to the maximum cumulative value from the reference group of samples of other organisms. 22. The method of claim 11, wherein the second threshold is a percentage, wherein comparing the second number of regions with the second threshold comprises dividing said second number of regions by the number of the second plurality of regions before comparing with said second threshold. 23. The method of claim 22, wherein said number of the second set of areas corresponds to the entire second set of areas. 24. The method according to claim 11, characterized in that the first set of areas coincides with the second set of areas. 25. The method according to claim 11, characterized in that said biological sample is taken before treatment, and further comprising repeating the method according to claim 11 for another biological sample taken after treatment, to obtain: the next first number of regions for which the presence of a deletion or amplification was determined; and the next second number of regions that were determined to have a methylation level of the region in excess of the respective limit value for the region; comparing the first amount with the next first amount and the second amount with the next second amount to determine the prognosis for said organism. 26. The method of claim 25, wherein comparing the first quantity with the next first quantity and the second quantity with the next second quantity to determine a prognosis for said organism comprises determining a first difference between the first quantity and the next first quantity; comparing said first difference with one or more thresholds for the first difference; determining a second difference between the second quantity and the next second quantity; and comparing the second difference with one or more second difference thresholds. 27. The method of claim 26, wherein making a worse prediction if the first difference is below one of the first difference thresholds than when said first difference is above one of the first difference thresholds, and making a worse prediction if the second difference is below one of the second difference thresholds than if said second difference is above one of the second difference thresholds. 28. The method of claim 25, wherein said treatment is immunotherapy, surgery, radiation therapy, chemotherapy, antibody-based therapy, epigenetic therapy, or targeted therapy. 29. The method according to claim 1, characterized in that the first limit value is a specified distance from the reference level of methylation, established from a biological sample obtained from a healthy organism. 30. The method of claim 29, wherein said distance is the specified number of standard deviations from the reference methylation level. 31. The method according to claim 1, characterized in that the first limit value is set according to the reference level of methylation, determined from the previous biological sample from the specified organism, obtained earlier than the biological sample under study. - 90 042157 32. The method according to claim 1, where the method is preceded by determining the fractional concentration of tumor DNA in a biological sample; determining whether the fractional concentration of tumor DNA in the biological sample exceeds a minimum value; and in the event that the fractional concentration does not exceed the minimum value, marking the specified biological sample as unsuitable for analysis. 33. The method of claim 32, wherein the minimum value is determined based on the expected difference between the methylation levels for the tumor and the reference methylation level. 34. The method according to claim 1, where the calculation of the first level of methylation also includes measuring the size of DNA molecules located at multiple sites; and normalizing the first methylation level based on the measured DNA molecular sizes before comparing the first methylation level with the first limit value. 35. The method according to p. 34, characterized in that the normalization of the first level of methylation based on the measured sizes includes the selection of DNA molecules having a first size, where the first size is predetermined; and using the selected DNA molecules to calculate the first level of methylation, the first limit corresponds to the first size. 36. The method of claim 35, wherein the first dimension is a range of lengths. 37. The method according to claim 35, characterized in that the DNA molecules are selected based on the physical separation of the DNA molecules by size. 38. The method of claim 35, wherein performing methylation-sensitive massive parallel sequencing yields a sequence pair for each DNA molecule, and selecting first size DNA molecules comprises determining the size of the DNA molecules by comparing the sequence pairs of the plurality of DNA molecules with a reference genome; and selecting DNA molecules having a first size. 39. The method of claim 34, wherein normalizing the first methylation level based on measured sizes comprises determining a functional relationship between size and methylation levels, wherein the functional relationship is determined by measuring methylation levels of DNA molecules of different sizes and plotting the measured values; and using said functional relationship to normalize the first level of methylation, where said functional relationship provides scaling values corresponding to the respective sizes. 40. The method according to item 39, where the use of functional dependence for normalization of the first level of methylation includes the calculation of the average size corresponding to the DNA molecules used to calculate the first level of methylation; and multiplying said first methylation level by the appropriate scaling value. 41. The method of claim 39, wherein the use of a functional relationship to normalize the first methylation level includes, for each of the plurality of sites: for each of the DNA molecules located at the specified site: obtaining the appropriate size of the DNA molecule at the specified site; and using a scaling value corresponding to said respective size to normalize the contribution of a DNA molecule to the corresponding number of DNA molecules that are methylated at said site. 42. The method of claim 1, wherein the plurality of sites includes CpG sites, wherein said CpG sites are organized into a plurality of CpG islands, each CpG island comprising one or more CpG sites, wherein the first level of methylation corresponds to to the first CpG island of the set of CpG islands. 43. The method of claim 42, wherein each of the CpG islands is selected as having a mean methylation density that is below a desired percentage in a reference group of samples of other organisms, and wherein each of the CpG islands is selected as having a coefficient of variation for mean methylation density in reference group below the desired percentage of multiple CpG islands. 44. The method of claim 42, further comprising, for each of the CpG islands: determining whether the CpG island is hypermethylated relative to a reference group of samples from other organisms by comparing the methylation level of the CpG island with an appropriate cut-off value; - 91 042157 for each of the hypermethylated CpG islands: determining the appropriate methylation density; calculating a cumulative index for the respective methylation densities; and comparing said cumulative score with the cumulative limit value to determine a first classification. 45. The method according to claim 44, characterized in that the calculation of the cumulative index for the respective methylation densities includes for each of the hypermethylated CpG islands: calculating the corresponding difference between the corresponding methylation density and the reference density; and calculating the corresponding probability, which is the probability of obtaining the corresponding difference by chance according to the statistical distribution; and applying the sum of the respective probabilities to determine the cumulative score. 46. The method according to claim 45, characterized in that the cumulative indicator is determined by: for each hypermethylated CpG island of hypermethylated CpG islands: calculating the logarithm of the corresponding probability to obtain the corresponding logarithmic result; and calculating the sum including the corresponding logarithmic results. 47. The method of claim 45, wherein each respective difference is normalized to a standard deviation associated with the reference density. 48. The method according to claim 44, characterized in that the cumulative limit value corresponds to the maximum cumulative indicator of the reference group. 49. The method of claim 44, wherein determining whether the first CpG island is hypermethylated comprises comparing the first methylation level with a first cut-off value and with a second cut-off value, wherein the first cut-off value corresponds to the mean methylation densities for the reference group plus 2 %, and the second limit value corresponds to three standard deviations plus the average value of the methylation densities for the reference group. 50. The method of claim 1, further comprising, for each of the first plurality of genome regions: determining the appropriate number of DNA molecules as originating from the specified region; calculating the corresponding normalized value for the corresponding amount; and comparing said corresponding normalized value with a reference value to determine if a deletion or amplification is observed in the respective region; determining a first set of regions, all of which have been determined to exhibit one of the deletion, amplification, or normal representation, wherein the first methylation level corresponds to the first set of regions; determining a second set of regions, all of which are determined to show one or more of the deletion, amplification, or normal presentation; and calculating a second methylation level based on the respective amounts of DNA molecules methylated at the sites of the second set of regions, wherein comparing the first methylation level with the first limit value includes calculating a parameter relating the first methylation level and the second methylation, where the parameter includes the ratio or difference between the first level methylation and a second level of methylation; and comparing said parameter with the first limit value. 51. The method of claim 50, wherein the first methylation level is a statistic reflecting the area methylation level calculated for each area of the first set of areas, and wherein the second methylation level is a statistic of area methylation levels calculated for each area of the second set of areas. 52. The method of claim 51, wherein said statistic is determined using Student's t-test, analysis of variance (ANOVA), or Kruskal-Wallis test. 53. The method of claim 50, wherein calculating said parameter comprises applying a probability distribution to the ratio or difference to obtain a probability that is the parameter. 54. The method according to any one of claims 1, 5-53, wherein said biological sample is obtained in such a way as to obtain nucleic acid molecules that are free from cells in the biological sample. 55. The method of claim 54, wherein the biological sample is selected from the group consisting of blood plasma and blood serum. - 92 042157 56. The method of claim 54, wherein the biological sample is obtained by centrifugation. 57. A computer-readable medium that stores a plurality of instructions that are needed when executing control commands by a computer system to implement the method described here, and these instructions include steps according to any of the items presented above. 58. A computer system comprising one or more processors and a computer readable medium storing a plurality of instructions that are needed when the computer system executes control commands to implement the method of any one of claims 1 and 5-56.