EA046998B1 - ANALYSIS OF THE TRANSCRIPTOME OF MATERIAL PLASMA USING MASSIVE PARALLEL RNA SEQUENCING - Google Patents

Description

Cross references to related applications

This application claims priority to U.S. Provisional Patent Application No. 61/770985, entitled Maternal Plasma Transcriptome Analysis Using Massively Parallel RNA Sequencing, filed February 28, 2013, which is incorporated herein by reference for all purposes.

Field of technology

The existence of extracellular pregnancy-specific RNA molecules has been reported in maternal plasma. They provided a non-invasive testing tool for fetal assessment and pregnancy monitoring using simply a maternal peripheral blood sample. To date, a number of promising options for prenatal diagnosis using maternal plasma RNA have been developed. Researchers have been actively seeking additional RNA markers to expand their application in different contexts of fetal disorders and pregnancy pathologies.

In theory, the most direct way to identify RNA markers is through direct profiling of extracellular RNA molecules in maternal plasma. However, this method is not simple because standard high-throughput screening technologies such as microarray analysis and serial analysis of gene expression (SAGE) have limited ability to detect the typically low concentrations of partially degraded extracellular RNA in maternal plasma. Therefore, most of the described screening strategies for RNA markers work indirectly, by comparing the expression profiles of the placenta and maternal blood cells. Only transcripts whose expression levels are significantly higher in the placenta than in maternal blood cells are subsequently examined in maternal plasma samples using highly sensitive but low throughput technologies such as real-time reverse transcriptase PCR (RT-PCR). The number of plasma RNA markers identified by this indirect method is still relatively limited. This may be because the tissue-based data acquisition strategy did not fully account for all biological factors that influence placental RNA levels in maternal plasma. In addition, transcripts that are expressed and released by non-placental tissues in response to pregnancy cannot be identified by this method. Thus, it would be highly desirable to obtain a sensitive and high-throughput technique that allows direct profiling of the maternal plasma transcriptome.

Likewise, direct profiling of plasma RNA may have application in other settings where there is a mixture of RNA molecules from two individuals, such as organ transplantation. The bloodstream of a transplant recipient contains nucleic acid molecules from both the donor and said recipient. Changes in the relative profile of RNA molecules originating from the donor or from the recipient may allow the detection of pathologies in the transplanted organ or recipient, such as graft rejection.

Brief description of the invention

Methods, systems, and devices are provided for diagnosing pregnancy-related disorders, determining allelic relationships, determining maternal or fetal contributions to circulating transcripts, and/or identifying maternal or fetal markers using a sample obtained from a pregnant female subject. In some embodiments, the sample is blood plasma containing a mixture of RNA molecules derived from the mother and the fetus.

RNA molecules are examined (eg, sequenced) and a set of reads is obtained, and the positions of said reads in a reference sequence are identified (eg, by sequence alignment). Informative loci are identified that are homozygous for the first allele in the mother or the fetus, and heterozygous for the first allele and the second allele in the other of the mother and fetus. Then, informative loci are selected, and reads located in the selected informative loci are additionally analyzed (for example, aligned to them). In some embodiments, the ratio of reads corresponding to the first alleles and the second alleles is calculated and compared to a threshold value for diagnosing a pregnancy-related disorder. In some embodiments, fetal-derived portions of the RNA in a sample are determined using reads located at selected maternal informative loci. In some embodiments, the ratio of reads corresponding to the first allele and the second allele is calculated for the individual selected informative locus, and the ratio is compared to a threshold value for determining the locus as a maternal or fetal marker.

The application of the proposed methods is not limited to prenatal diagnosis; they can be applied to any biological sample that contains a mixture of RNA molecules originating from two individuals. For example, a blood plasma sample obtained from an organ transplant recipient may be used. Transcripts expressed in the transplanted organ reflect the donor's genotype and are present at detectable levels in the recipient's blood. When reading these transcripts, informative loci can be identified in the composition of genes expressing

- 1,046,998 controlled by both donor and recipient, and the relative expression levels of alleles from each individual can be measured. Abnormal levels of expression of an allele derived only from the donor or only from the recipient can be used to diagnose transplant-related disorders.

Biological samples that are suitable for use in the proposed methods include blood, plasma, serum, urine, saliva and tissue samples. For example, fetal nucleic acids have been detected in the urine of pregnant women. The urine of kidney transplant recipients has been shown to contain cell-free nucleic acids and cells from the transplanted organ. Microchimerism has been observed under a variety of conditions. Microchimerism refers to the presence of a source of cells or nucleic acids from another individual in the body, including organs and tissues of a particular individual. Microchimerism has been observed in tissue biopsies of the thyroid gland, liver, spleen, skin, bone marrow and other tissues. Microchimerism can occur as a result of previous pregnancies or blood transfusions.

It is also proposed that the gene be used to diagnose a pregnancy-related disorder in a pregnant female subject. The expression level of the specified gene is compared with a control value determined from one or more other female subjects, each of whom is pregnant with a healthy fetus. Pregnancy-related disorders as defined herein include preeclampsia, intrauterine growth restriction, invasive placentation, and preterm birth. Other pregnancy-related disorders may be conditions that pose a risk of fetal death, such as hemolytic disease of the newborn, placental insufficiency, hydrops fetalis, and fetal malformations. Other pregnancy-related diseases may also include conditions that lead to complications during pregnancy, such as HELLP syndrome, systemic lupus erythematosus and other maternal immunological diseases.

Brief description of drawings

In fig. Figure 1 shows a positive relationship between the B/A allelic ratio and the relative expression levels of the corresponding genes in the placenta and maternal blood cells.

In fig. Figure 2 shows the relationship between the overall level of gene expression and the relative levels of gene expression in the placenta and maternal blood cells.

In fig. 3 graphically presents methods for diagnosing a pregnancy-related disorder.

In fig. Figures 4 and 5 show the generalized results of alignment of reads obtained from RNA sequencing.

In fig. Figure 6 shows RNA sequencing data from blood plasma for two individuals. In fig. Figure 6A shows the percentage distribution of GCs in sequence reads when sequencing the M9356P (with Ribo-Zero Gold pretreatment) and M9415P (without Ribo-Zero Gold pretreatment) RNA libraries. In fig. Figure 6B shows the correlation of the gene expression profiles of M9356P and M9415P.

In fig. Figure 7 shows RNA-SNP allelic relationships and the dependence of total concentrations of various RNA transcripts in maternal plasma on relative tissue expression levels (placenta/blood cells).

In fig. Figure 8 shows RNA-SNP allelic relationships and the dependence of total plasma levels of different genes containing informative SNPs on relative tissue expression levels (blood cells/placenta).

In fig. Figure 9 shows RNA-SNP allelic relationships for specific fetal SNP alleles and shows the expression levels of RNA transcripts containing these alleles in maternal plasma.

In fig. Figure 10 shows PHK-SNP allelic relationships for specific maternal alleles. SNPs show the expression levels of RNA transcripts containing these alleles in maternal plasma.

In fig. Figure 11 presents data for specific maternal SNP alleles in subjects with preeclampsia and control subjects. In fig. 11A shows RNA-SNP allelic relationships for case 5641 (late-onset third trimester preeclampsia) and case 7171 (control) for informative SNPs containing specific maternal SNP alleles. In fig. 11B shows the fold difference in allelic ratios and plasma RNA-SNP levels for case 5641 and case 7171 (control) for informative SNPs containing specific maternal SNP alleles.

In fig. Figure 12 presents data for specific maternal SNP alleles in subjects with preeclampsia and control subjects. In fig. 12A shows the RNA-SNP allelic relationships for case 5641 (late-onset third trimester preeclampsia) and case 9356 (control) for informative SNPs containing specific maternal SNP alleles. In fig. 12B shows the fold difference in allelic ratios and plasma levels of RNA-SNPs for case 5641 and case 9356 for informative SNPs containing specific maternal SNP alleles.

In fig. 13 shows the data from FIG. 11 in table form.

In fig. 14 shows the data from FIG. 12 in table form.

In fig. 15 graphically presents methods of determination in a sample obtained from a sub

- 2,046,998 ects - of a woman pregnant with a fetus, part of the RNA derived from the fetus.

In fig. Figure 16 shows the results of the analysis of informative SNPs in the maternal plasma transcriptome.

In fig. Figure 17 presents the relative contributions of the fetus and mother without selection for allele-specific expression in maternal plasma samples in late pregnancy.

In fig. Figure 18 shows the fetal and maternal contributions to the maternal plasma transcriptome. In fig. Figure 18A shows the proportions of fetal- and maternal-derived transcripts in maternal plasma during early and late pregnancy. In fig. 18B shows the calculated number of alleles and the proportion of fetal alleles before and after birth.

In fig. Figure 19 shows a comparison of the specific contributions of the fetus and mother for one RNA-SNP for a pregnancy with preeclampsia and a control pregnancy.

In fig. 20 graphically presents methods for determining a genomic locus as a maternal or fetal marker.

In fig. 21 provides a list of genes associated with pregnancy.

In fig. 22 shows lower expression levels of pregnancy-related genes in postpartum subjects compared to prepartum subjects.

In fig. Figure 23 shows hierarchical clustering of plasma samples using 131 pregnancy-related genes.

In fig. Figure 24 shows the expression levels of 131 pregnancy-related genes in the placenta, maternal blood cells, and maternal plasma. In fig. 24A shows a gene expression intensity map (log ₂ (transcription level)) for placenta and maternal blood cells, as well as antepartum and postpartum maternal plasma from two late-term pregnancies. The 131 pregnancy-related genes were predominantly expressed in the placenta. In fig. 24B shows that the expression levels of the 131 genes in the placenta and plasma were positively correlated (P < 0.05, Spearman correlation).

In fig. 25 provides a comparison of fetal and maternal plasma RNA contributions for genes identified by plasma-based and tissue-based methods.

In fig. 26 lists 98 preeclampsia-associated maternal plasma RNAs identified by RNA-Seq. Expression levels in plasma obtained from uncomplicated pregnant women and women who developed preeclampsia are shown.

In fig. Figure 27 shows the number of reads during RNA sequencing in the sites of single nucleotide polymorphisms (SNPs) of PAPPA RNA for case 9415. Allele A is a common allele, and allele G is a fetus-specific allele.

In fig. Figure 28 shows the number of reads during RNA sequencing at the sites of single nucleotide polymorphisms (SNP) of H19 RNA for case 9415.

In fig. Figure 29 shows determination of maternal H19 methylation status using methylation-sensitive enzymatic digestion.

In fig. 30 is a block diagram of an example computer system 3000 suitable for use in the system and methods of embodiments of the present invention.

Detailed Description of the Invention

I. Introduction

The presence of cell-free fetal RNA in maternal plasma was described more than a decade ago. Since this discovery, many studies have been carried out to detect circulating RNAs of fetal and placental origin in maternal plasma. Interestingly, the expression levels of placenta-specific transcripts in plasma were found to be positively correlated with those in placental tissues, highlighting the clinical relevance of plasma RNA analysis as a non-invasive tool for monitoring the health and development of the placenta or fetus. Thus, maternal circulating RNA testing has found clinical application in pregnancy- and placenta-related disorders such as preeclampsia, intrauterine growth restriction, and preterm birth, as well as for non-invasive testing of fetal chromosomal aneuploidies. Such developments highlight the potential utility of RNA biomarkers for the molecular assessment of prenatal disorders.

Despite the promise of plasma RNA analysis for prenatal testing, to date the number of reliably validated pregnancy- or placenta-associated transcripts in maternal plasma is still limited. Therefore, RNA biomarkers in plasma were examined using reverse transcriptase (RT)-PCR, a sensitive method that typically targets a relatively small number of RNA species per assay. Notably, these studies have primarily focused on genes whose expression levels are relatively high in the placenta compared to maternal blood cells, based on the premise that pregnancy-related genes are primarily placenta-derived. Such approaches have provided the identification of pregnancy-associated target RNAs with high levels of expression in the placenta, but other important targets may have been missed. In addition, given the low concentrations and low stability of plasma RNA ^9.1 5, standard high-throughput methods such as serial gene expression analysis and microarray

- 3 046998 ric analysis are not reliable methods for directly studying the plasma transcriptome.

These technical limitations in plasma RNA analysis can potentially be overcome by using massively parallel sequencing (MPS) for RNA analysis, namely RNA sequencing (RNA-seq). Due to its increased sensitivity and wide dynamic range, RNA-seq has been used to study gene expression in many human tissues, including the placenta. ¹⁹ The advantage of MPS has also been confirmed by its applicability in direct profiling of plasma microRNAs from healthy individuals and pregnant women. However, the full spectrum of the plasma transcriptome is still unclear, possibly due to the lower stability of long RNAs in plasma compared to short miRNAs.

In the present study, the authors showed that transcripts of fetal and maternal origin can be detected in maternal plasma and that their relative contributions can be assessed using RNA-seq by examining specific fetal and maternal single nucleotide polymorphisms (SNPs). In addition, maternal plasma can monitor allele-specific expression patterns in the placenta. The authors also demonstrated that pregnancy-related transcripts could be identified by directly examining maternal plasma before and after delivery.

II. Method for diagnosing pregnancy-related disorder using SNP allele relationships

Gene expression profiling is suitable for use in determining the disease status of an individual. Embodiments of the present invention include methods for analyzing the expression profile of an individual by analyzing a mixture of RNA molecules from two different individuals. These methods use the relative abundance of alleles specific to one individual and alleles common to two individuals. Based on the relative abundances of these common and individual-specific alleles, the gene expression profile of the two individuals can be determined. One possible application of the proposed methods is to analyze the fetal gene expression profile by analyzing RNA in a maternal plasma sample that contains RNA from both the fetus and the pregnant woman. Another application is the analysis of the gene expression profile of donor-derived RNA in a plasma sample obtained from a transplant recipient containing RNA from both the donor and the recipient.

In a mixture containing RNA from two individuals, the expression profile of each individual cannot be determined from analysis of the total amount of different RNA transcripts in the mixture, since it is difficult to determine the relative contributions of each individual to the total RNA. The relative abundance of different RNA transcripts can be used to monitor an individual's health or to determine a disease state.

According to the present method, the genotype of the two individuals can first be determined by direct genotyping of the individuals or by family analysis. For example, if the parents' genotypes are AA and TT, the fetus will have the AT genotype.

The following hypothetical example illustrates the principle of this method. For each of the genes of interest, there is a SNP within the coding region, so polymorphism can be observed in the RNA transcripts of different genes. Let us assume that the genotypes of the fetus and the pregnant woman are AB and AA, respectively. Accordingly, allele B will be specific to the fetus, and allele A will be common to the fetus and mother. The relative expression levels of different genes in the placenta and maternal blood cells correspond to those given in Table. 1.

In this example, the authors assumed that the placenta and maternal blood cells contribute to maternal plasma in relatively equal proportions, representing 2% of each of their RNA transcripts. In other words, if the placenta has a gene expression level of 10,000, the contribution to the maternal plasma will be 200 RNA transcripts of the specified gene. As can be seen from Fig. 1, there is a well-defined positive relationship between the B/A allelic ratio and the relative levels of expression of the corresponding genes in the placenta and maternal blood cells. In contrast, the overall level of gene expression is influenced by fluctuations in expression in maternal blood cells and therefore correlates less well with expression in the placenta (Fig. 2).

- 4 046998

Gene Maternal genotype Fetal genotype Relative expression level Relative expression (placenta/maternal blood cells) Mother plasma in maternal blood cells in the placenta Relative maternal/fetal RNA abundance Total concentration RNA RNA concentration B/A ratio with allele A with allele B 1 AA AB 8000 20000 2.5 1:1 560 360 200 0.56 2 AA AB 6000 10000 1.67 1:1 320 220 100 0.45 3 AA AB 10000 5000 0.5 1:1 300 250 50 0.2 4 AA AB 14000 1000 0.071 1:1 300 290 10 0.03

Another advantage of using allelic relationship analysis is that the level of expression of a particular gene in the placenta is normalized to the level of expression of maternal blood cells. Since expression levels can vary significantly between genes and different samples, this normalization makes comparisons between different genes more reliable and avoids the need for comparison to a reference genome, such as a housekeeping genome. In other words, in standard gene expression analysis, the expression level of a gene in a sample is either measured in relation to a housekeeping gene or expressed as a proportion of total RNA in the sample. To identify aberrant gene expression, the reported relative values for the test sample are then typically compared with those for control samples.

Here, the authors propose a new approach that uses the relative amounts contributed by the fetus, normalized by the maternal contribution for the same gene, as a way to determine gene expression profiles. In pathological conditions such as preeclampsia, preterm birth, maternal diseases such as systemic lupus erythematosus, in which the expression of the specified gene in the placenta or maternal organs is altered, the fetal/maternal ratio for the specified gene will be altered compared to pregnancies without such pathological conditions. This approach can be implemented using a specific fetal allele SNP relative to the common allele in maternal plasma RNA transcripts.

This approach can also be implemented using the ratio of specific maternal allele SNPs and common alleles in RNA transcripts from maternal plasma. Pathological conditions can be identified if one or more of these allelic relationships for one or more gene RNA transcripts are altered from those expected for a non-pathological condition. The pattern of such allelic relationships at a gene locus or multiple gene loci can be used to identify pathological conditions. Non-pathological conditions may be represented by allelic ratios in normal pregnancies, in samples obtained before or after testing when the pregnancy is no longer affected by the pathological condition, or by existing data previously obtained from normal pregnancies, e.g. previously obtained reference range.

Pregnancy-related disorders that can be diagnosed using the proposed methods include any disorder characterized by abnormal relative levels of gene expression in maternal and fetal tissue. These disorders include, but are not limited to, preeclampsia, intrauterine growth restriction, invasive placentation, preterm birth, hemolytic disease of the newborn, placental insufficiency, hydrops fetalis, fetal malformations, HELLP syndrome, systemic lupus erythematosus and other maternal immunological diseases. These methods make it possible to distinguish between RNA molecules contributed by the mother and the fetus in a sample that contains a mixture of such molecules. Accordingly, these methods make it possible to identify changes in the contribution of one individual (ie, mother or fetus) to a specified mixture at a particular locus or for a particular gene, even if the contribution of another individual does not change or changes in the opposite way. Such changes cannot be easily determined by measuring the overall level of expression of the gene without taking into account the tissue or individual origin. Pregnancy-related disorders as discussed herein are not characterized by fetal chromosomal abnormalities such as aneuploidy.

The proposed methods can also be implemented without previously obtained information about the genotype of the placenta. For example, the genotype of a pregnant woman can be determined by direct genotyping of her blood cells. The RNA transcripts of the plasma samples can then be analyzed, for example, but not limited to, massively parallel sequencing. RNA transcripts in which two different alleles are observed can be identified. In the specified set of transcripts, if the mother is homozygous, this will indicate heterozygosity of the fetus for the specific fetal allele and the maternal allele. In this case, analysis of the allelic ratio of RNA can be carried out without previously obtained information about the genotype

- 5 046998 fetus (or placenta). Additionally, cases where the mother is heterozygous and the fetus is homozygous can be identified by deviation from the 1:1 ratio. More information on such techniques is provided in US Patent 8467976.

A. Diagnostic method

A method 300 for diagnosing a pregnancy-related disorder, in accordance with some embodiments, is illustrated in FIG. 3. This method uses a sample obtained from a subject - a woman pregnant with a fetus. This sample can be obtained from the mother's blood plasma; it contains a mixture of RNA molecules derived from both the mother and the fetus. The sample may be obtained by a suitable method from a female subject at any stage of pregnancy. For example, subjects may be pregnant women with singleton pregnancies in the first, second, and third trimester. Pregnant women in the first and second trimester may be classified as early pregnancy cases; pregnant women in the third trimester may be classified as late-term pregnancies. In some embodiments, samples may also be collected after birth of the fetus, or from the same subject before and after delivery, and samples from non-pregnant women may be used as controls. Samples can be obtained at any time after delivery (eg, 24 hours).

Said sample may be obtained from peripheral blood. The maternal blood sample may be processed by a suitable method, such as centrifugation, to separate the blood cells from the plasma. Stabilizers may be added to the blood sample or parts thereof, and the sample may be stored until use. In some embodiments, samples of fetal tissue, such as chorionic villi, amniotic fluid, or placental tissue, are also obtained. Fetal tissue can be used to determine fetal genotypes as described below. Fetal tissue can be obtained before or after birth.

Once a sample is obtained, RNA or DNA can then be extracted from maternal blood and plasma cells, as well as from any fetal tissue, such as a placenta sample. Placental and blood cell RNA samples can be pretreated, for example, using the Ribo-Zero Gold Kit (Epicentre) to remove ribosomal RNA (rRNA) before preparing the library for sequencing.

At step 301 of the method, a set of reads is obtained. These reads are obtained by analyzing RNA molecules obtained from the specified sample. In various embodiments, these reads can be obtained by sequencing, digital PCR, RT-PCR, and mass spectrometry. Although the discussion is focused on sequencing, aspects of the present description also apply to other techniques for obtaining reads. For example, digital PCR (including microfluidic and droplet PCR) can provide information about different alleles at a particular locus through the use of probes (including primers) targeting each allele. These probes may contain marks that allow them to be distinguished from each other, for example, by coloring them. Within the same experiment or a different experiment (e.g., on a separate slide or chip), probes with different labels can target different loci. Tag detection reflects the presence and/or quantity of RNA or cDNA molecules being amplified and corresponds to reads of the RNA or cDNA molecules, such reads providing information regarding the sequence of the RNA or cDNA molecules at loci corresponding to the probes. The description including sequencing reads applies equally to reads obtained using any suitable technique, including non-sequencing techniques.

Sequencing can be carried out in a suitable manner using any available technology. Examples of nucleic acid sequencing technologies and methods include massively parallel sequencing, next-generation sequencing, whole-genome sequencing, exome sequencing, sequencing with or without target enrichment, sequencing by synthesis (e.g., amplification sequencing, clonal amplification, bridging amplification, sequencing with reversible terminators) , ligation sequencing, hybridization sequencing (eg, microarray sequencing), single-molecule sequencing, real-time sequencing, nanopore sequencing, pyrosequencing, semiconductor sequencing, mass spectrometric sequencing, shotgun sequencing, and Sanger sequencing. In some embodiments, sequencing is performed using massively parallel techniques on a cDNA library derived from the RNA in the sample. cDNA libraries can be synthesized by a suitable method, for example, using the mRNA-Seq Sample Preparation Kit (Illumina), according to the manufacturer's instructions or with minor modifications. In some embodiments, a 5-fold diluted fragment of Klenow DNA polymerase is used for the plasma cDNA end repair step. The QIAquick PCR Purification Kit and QIAquick MinElute Kit (Qiagen) can be used to purify end-reduced and andenylated products, respectively. In some embodiments, 10-fold diluted paired-end adapters are used for plasma cDNA samples or two rounds of product purification are performed

- 6 046998 with ligated adapters using AMPure XP granules (Agencourt). The cDNA library can be sequenced with read lengths of 75 bp. in paired-end format on a HiSeq 2000 instrument (Illumina), or using other instrument formats.

At step 302, the position of the reads in the reference sequence is determined. When using PCR-based techniques, the location can be determined, for example, by matching the colors of a detectable label on a probe or primer with the sequence for which said probe or primer is specific. When using sequencing techniques, the position can be determined by the alignment of the sequence read with a reference sequence. Sequence alignment can be accomplished using a computer system. Any bioinformatics algorithm (e.g., removal of highly duplicated reads and low quality reads), data alignment, and/or data normalization can be used to preprocess raw data. RNA transcription levels in maternal blood cells, placental tissues, and maternal plasma can be calculated as fragments per kb per million exonic reads (FPKM). Data preprocessing and alignment may be performed to determine allelic relationships, but data normalization is not required. Any reference sequence (for example, the human reference genome hg19) and any algorithm can be used for alignment.

In one example, steps 301 and 302, after removing duplicated and rRNA reads, resulted in an average of 3 million analyzable reads for each non-pregnant plasma sample; obtained an average of 12 million reads suitable for analysis for each plasma sample from pregnant women. For tissue RNA-seq, an average of 173 million and 41 million analyzable reads were obtained per sample for placenta and blood cells, respectively. RNA-seq alignment statistics are summarized in Fig. 4 and 5. The GC content of the sequencing reads was also examined in all samples (Fig. 6A and Table 2).

Table 2. Proportion of GC-rich sequencing reads in RNA-seq libraries

Sample type Sample Treatment GC-rich reads (% raw reads) ^a Plasma М9356Р rRNA depletion 49.28% М9415Р No 0.00% Placenta N9356 rRNA depletion 0.15% Blood cells M9356W rRNA depletion 0.30%

a Raw reads were aligned to the de novo assembled GC-rich sequence using BLAST (NCBI).

At step 303, one or more informative loci are identified. This stage allows you to determine genotypes or draw conclusions regarding the genotypes of the female subject and the fetus at loci in the genome, and compare these genotypes. A locus is considered informative if it is homozygous in the first object for the corresponding first allele (for example, AA), and heterozygous in the second object for the corresponding first allele and the corresponding second allele (for example, AB). Said first object may be a pregnant woman or a fetus subject, and said second object may be another subject of a pregnant woman and a fetus. In other words, one individual (either the mother or the fetus) is homozygous and the other individual is heterozygous for each informative locus.

Informative loci can be further classified according to which individual is heterozygous and the sole source of one allele. The locus is considered an informative maternal locus, or, equivalently, a specific maternal (maternal specific) locus, if the fetus is homozygous and the subject, the pregnant woman, is heterozygous. A locus is considered a fetal informative locus, or, equivalently, a fetal-specific locus, if the subject, the pregnant woman, is homozygous and the fetus is heterozygous. At step 303, all identified informative loci are either maternal-specific or fetal-specific, since the same individual is used as the second object for all of the identified loci.

An informative locus can be represented by a single nucleotide polymorphism, or SNP, when the first allele and the second allele differ in a single nucleotide. An informative locus may also be represented by a short insertion or deletion, where one or more nucleotides are added or deleted in one allele relative to another allele.

In some embodiments, the genotype of a pregnant subject at one or more informative loci, or each informative locus, is determined by sequencing genomic DNA obtained from maternal tissue. Said maternal tissue may be maternal blood cells or any other type of tissue. According to some

- 7 046998 embodiments, the genotype of the fetus at one or more informative loci, or at each informative locus, is determined by sequencing genomic DNA obtained from fetal tissue, such as the placenta, chorionic villi, amniotic fluid. If a less invasive method is preferred, fetal DNA for sequencing can also be obtained from a maternal blood sample. The source of such fetal DNA may be the same sample as that used to obtain the RNA molecules for sequencing, or a different sample may be used. It should be kept in mind that fetal informative loci can be identified without direct genotyping of the fetus. For example, if a pregnant subject is determined to be homozygous for the first allele at one locus and RNA sequencing indicates the presence of the second allele in a mixture of maternal and fetal RNA, the fetus may be considered heterozygous at that locus.

In some embodiments, both mother and fetus are genotyped using massively parallel sequencing techniques to identify informative loci. Sequencing can be performed on exome-enriched samples of genomic DNA from maternal and placental blood cells. Genomic DNA can be extracted from placental tissues and maternal blood cells using the QIAamp DNA Kit and QIAamp Blood Kit (both from Qiagen), respectively, following the manufacturer's instructions. Exome enrichment and sequencing library generation can be performed using the TruSeq Exome Enrichment Kit (Illumina) according to the manufacturer's protocol. Libraries can be sequenced with read lengths of 75 bp. in PE sequencing format on a HiSeq 2000 instrument (Illumina).

At step 304, one or more informative loci are selected. As used herein, selection means selecting certain informative loci from those identified at step 303 for further analysis. The selected informative locus is the locus selected in this manner. The selection can be made based on one or more criteria. One such criterion is where the specified informative loci are located in the genome or on a reference sequence. In some embodiments, only loci located in an exon or expressed region of the reference sequence are further analyzed.

Another selection criterion is the number of sequencing reads that align to the specified locus and contain each allele, from those reads that were obtained in step 301 and aligned with the reference sequence in step 302. In some embodiments, at least at least a first predetermined number of sequencing reads containing the first allele, and/or a second predetermined number of sequencing reads containing the second allele. This predetermined number may be 1, 2, or more, and may correspond to the required read quality or sequencing depth. As a result of the selection, one or more selected informative loci are identified.

In some embodiments, only informative loci representing SNPs are considered for selection. According to one example implementation of this method, approximately 1 million SNPs of the dbSNP NCBI database, version 135, all of which are located in exons, were examined, and informative ones were identified from these 1 million SNPs using maternal and fetal genotypes. To illustrate informative SNPs, A was specified as the common allele and B as the maternal specific or fetal specific allele. Informative SNPs included SNPs with a maternal allele-specific SNP, namely fetal AA and maternal AB at a specific locus, and SNPs with a fetal allele-specific SNP, namely maternal AA and fetal AB, at a specific locus. Genotypes were predicted using our own bioinformatics algorithm. To perform the selection, only informative SNPs that had at least one counted read for both allele A and allele B, respectively, were included in the analysis.

At steps 305 and 306, the sequencing reads aligned to each selected informative locus are counted and sorted based on which allele they contain. For each selected informative locus, two numbers are determined: the first number, which is the number of sequencing reads that align to the specified locus and contain the corresponding first allele (ie, the common allele); and a second number, which is the number of sequencing reads that align to the specified locus and contain the corresponding second allele (ie, maternal specific or fetal specific allele). The specified first number and second number can be calculated in a suitable manner. Said first number plus said second number for each locus gives the total number of sequencing reads aligned to said locus. The ratio of said first number to said second number for a particular selected informative locus may be referred to as an allelic ratio.

At steps 307 and 308, the number of first alleles and second alleles at all selected informative loci is summed. At step 307, the first sum of the first numbers is calculated. The first sum reflects the total number of sequencing reads containing the corresponding first alleles (i.e., common alleles) for the selected informative loci. At step 308, a second sum of the second numbers is calculated. The second sum reflects the total number of sequencing reads containing the corresponding second alleles (i.e.

- 8 046998 specific maternal or specific fetal alleles) for selected informative loci. These sums may be calculated in a suitable manner using all of the selected information loci identified at step 304, or a subset thereof. In some embodiments, these amounts are weighted. For example, the contributions to the indicated sums of loci from a particular chromosome can be scaled, down or up, by multiplying the first numbers and second numbers for the indicated loci by a scalar.

At step 309, the ratio of the first amount and the second amount is calculated. This ratio reflects the relative number of sequencing reads for the first alleles and the second alleles combined across the selected informative loci. In some embodiments, the ratio is calculated by simply dividing the first amount by the second amount. This calculation gives the number of sequencing reads for common alleles as a set of sequencing reads for specific maternal or specific fetal alleles. Alternatively, said ratio may be calculated by dividing the second amount by the sum of said first amount and second amount. In this case, this ratio gives the number of sequencing reads for specific maternal or specific fetal alleles as a proportion of all sequencing reads. Other methods for calculating this ratio will be apparent to one skilled in the art. The first sum and the second sum from which the ratio is calculated can be used as a proxy for the amounts of RNA (ie transcripts) present in the sample that are derived from common and specific alleles for the selected informative loci.

At step 310, the ratio calculated at step 309 is compared to a threshold value to determine whether the fetus, mother, or pregnancy has a pregnancy-related disorder. In various embodiments, the threshold may be determined from one or more samples obtained from pregnant women subjects who do not have a pregnancy-related disorder (i.e., control subjects) and/or determined from one or more samples, obtained from subjects who are pregnant women who have a pregnancy-related disorder. In some embodiments, the threshold value is determined by implementing the same method as described above on samples from control subjects and examining the same or overlapping set of selected informative loci. The threshold value can be set as a value between the expected values for pregnancies without the disorder and the expected values for pregnancies with the disorder. The threshold value may be based on a statistical difference from the normal value.

In some embodiments, the ratio calculated for the pregnant subject (at step 309) uses selected information loci from a particular set of genes, and the ratio (namely, the threshold value) is calculated for control subjects using the selected information loci in the genes, some or all of which match. If the calculated ratio for a pregnant subject is based on specific maternal alleles, then the calculated threshold for control subjects may also be based on specific maternal alleles. Likewise, both the ratio and the cutoff may be based on specific fetal alleles.

Comparison of the ratio with a threshold value can be carried out in a suitable manner. For example, a disorder may be diagnosed when a specified ratio is greater than a threshold value or less than a threshold value by any amount or by an amount greater than a certain limit. The comparison may include calculating a difference or calculating another ratio between the ratio for the pregnant woman subject and the threshold value. In some embodiments, the comparison includes assessing whether there is a significant difference in the relative expression levels of common and non-common alleles for certain genes between a pregnant subject and control subjects.

If necessary, this method can also be used to evaluate the portion of RNA in the sample that is of maternal or fetal origin. This estimate is obtained by multiplying the ratio calculated at step 309 by a scalar. This scalar reflects the overall level of expression at selected informative loci relative to the expression of second alleles in a heterozygous individual.

The following example illustrates the use of this scalar. In the event that the second alleles counted by this method are fetal specific, the ratio calculated at step 309 may yield the number of sequencing reads for the fetal specific alleles, represented as a proportion of all sequencing reads for the selected informative loci. Among the sequencing reads containing a common allele, some were contributed by the mother and some were contributed by the fetus. If the relative expression levels of fetal specific and fetal common alleles are known or can be estimated, the ratio can be scaled to provide an estimate of the specific fetal contribution to all sequencing reads for selected informative loci.

In some embodiments, the fetal and common alleles are expressed symmetrically at most loci and the scalar is assigned a value close to 2. In other embodiments

- 9 046998 implementations the specified scalar is not equal to 2 and takes into account asymmetric gene expression ²⁴ . The relative maternal contribution to the sequencing reads is then obtained by subtracting the relative fetal contribution from unity. Similar calculations can be performed if the second alleles in the selected informative loci are specific to the mother.

B. Examples

1. Correlation of allelic ratios and plasma concentrations

The expression profile of a maternal plasma sample (gestational age 37 2/7) was analyzed and compared with corresponding samples of placenta and maternal blood cells from a pregnant woman. Massively parallel RNA sequencing was performed for each of these samples using the Illumina HiSeq 2000 instrument. Placenta and maternal blood cells were genotyped using massively parallel exome sequencing. The authors examined approximately 1 million SNPs from the NCBI dbSNP database, ver. 135 located in exons, and classified informative SNPs for which the mother is homozygous (AA genotype) and the fetus is heterozygous (AB genotype). Accordingly, allele A will be an allele common to the mother and fetus, and allele B will be specific to the fetus.

RNA-SNP allelic ratios and total concentrations of different RNA transcripts in maternal plasma are compared with relative tissue expression levels (placenta/blood cells) (Fig. 7). The authors observed that the RNA-SNP allelic ratio in plasma correlated well with the relative tissue expression level (Spearman R = 0.9731679, P = 1.126e-09). However, total plasma levels did not correlate with tissue expression (Spearman R = -0.7285714, P = 0.002927).

In addition to fetal expression analysis, this method can also be used for maternal expression profiling. In these circumstances, informative SNPs can be used for which the mother is heterozygous (AB genotype) and the fetus is homozygous (AA genotype). In such a case, the RNA-SNP allelic ratio can be calculated by dividing the counted number of specific maternal alleles by the counted number of common alleles. In fig. Figure 8 shows the relationship between allelic RNA-SNP ratios and total plasma levels of various genes containing informative SNPs and the relative level of tissue expression (blood cells/placenta). The authors observed a strong positive association between tissue expression levels and RNA-SNP allelic ratios in maternal plasma (Spearman's R = 0.9386, P < 2.2e-16), but not total plasma transcription levels (Spearman's R = 0.0574 , P=0.7431).

The RNA transcripts, allelic ratios of the RNA SNPs they contain, and plasma levels are presented in the tables in Figs. 9 and 10.

2. Comparison of RNA-SNP allelic ratios in maternal plasma for preeclamptic and control cases

For the two cases with pregnant women subjects, an average of 117,901,334 raw fragments were obtained (Table 3A). The RNA-SNP allelic ratio of the subset of circulating transcripts carrying specific maternal SNP alleles was compared for a late-onset third trimester (PET) preeclampsia case, 5641, and a gestational age-matched control case, 7171. As can be seen in FIG. 11A, RNA-SNP allelic ratios for the indicated transcripts reflect different profiles in the PET case and the control case. Importantly, the profile of RNA-SNP allelic relationships differs from that of plasma transcription levels, as seen in Fig. 11B. This suggests that analysis of RNA-SNP allelic relationships in maternal plasma can be used as an additional indicator to better differentiate PET cases from normal pregnancies. The number of informative transcripts can be further increased by increasing the number of informative SNPs, as shown by comparing the PET case, 5641, and another normal pregnancy control case, 9356 (Fig. 12A and 12B). RNASNP allelic ratios and plasma levels of RNA transcripts presented in Fig. 11 and 12 are shown in tabular form in FIG. 13 and 14 respectively.

Table 3A. Summary of RNA-seq read alignment results

Sample type Sample Raw reads Preprocessed ³ reads (%) ^b Mapped reads ^with (%) ^d Mother plasma 5641 (Preeclampsia) 119 902 178 30 106 800 25.11% 4 001 376 13.29% Mother plasma 7171 (control) 115 900 490 29 160 920 25.16% 1 397 345 4.79%

A. Remaining reads after removing frequently duplicated reads. b.% raw reads.

c. See additional table. 2B with detailed explanation.

d. % of preprocessed reads.

- 10 046998

Table 3B. Summary of RNA-seq read alignment results

Sample type Sample Selected reads ^{a Reads} suitable for analysis Exon (%) Intron (%) Intergenic region (%) ^with Mother plasma 5641 (Preeclampsia) 2 579 163 1 422 213 44.87% 12.32% 42.80% Mother plasma 7171 (control) 790 735 606 610 37.69% 16.09% 46.22%

A. The selected reads mainly represent nuclear and mitochondrial rRNAs and tRNA b. Reads suitable for analysis = total number of mapped reads - selected reads. c. % of reads aligned to regions outside exons and introns of reference genes.

III. Method for determining fetal or maternal RNA contribution

A. Method

In fig. 15 illustrates a method 1500 for detecting a portion of fetal-derived RNA in a sample obtained from a fetal subject. The specified sample can be obtained and prepared according to the above description for method 300.

At step 1501, a set of sequencing reads is obtained. At step 1502, said sequencing reads are aligned to a reference sequence. These steps may be carried out as described above for steps 301 and 302.

At step 1503, one or more maternal informative loci are identified. Each informative maternal locus is homozygous in the fetus for the corresponding first allele and heterozygous in the subject - a pregnant woman for the corresponding first allele and the corresponding second allele. The first allele, accordingly, is common to the mother and fetus, and the second allele is specific to the mother. Informative maternal loci may be identified as described above at step 303.

At step 1504, maternal informative loci are selected according to step 304, and one or more selected maternal informative loci are identified. Sequencing reads are processed at the level of individual informative loci selected at stages 1505-1511. At steps 1505 and 1506, a first number and a second number for each selected informative maternal locus are determined. The first number is determined at step 1505 as the number of sequencing reads that align to the specified locus and contain the corresponding first allele. The second number is determined at step 1506 as the number of sequencing reads that align to the specified locus and contain the corresponding second allele. Steps 1505 and 1506 may be performed in a manner similar to steps 305 and 306 as described above.

At step 1507, the sum of said first number and said second number is calculated for each informative maternal locus. The specified amount may be equivalent to the total number of sequencing reads aligned to the specified locus. At step 1508, the maternal ratio is calculated by dividing the second number by the specified amount. Maternal ratio gives the proportion of all sequencing reads aligned to a particular locus that contain a second (maternal-specific) allele.

At step 1509, a scalar is determined. The specified scalar reflects the overall level of expression in the selected informative maternal locus relative to the expression of the corresponding second allele in the subject - a pregnant woman. In some embodiments, if the expression of two alleles in a pregnant subject is known or the expression is assumed to be symmetrical, the scalar is approximately equal to 2. In other embodiments, the value of the scalar may be different from 2 to account for asymmetric gene expression. Since most loci are expressed symmetrically, the symmetry assumption applies.

At step 1510, the maternal ratio is multiplied by a scalar to obtain the maternal contribution. Maternal contribution reflects, at a selected informative maternal locus, the proportion of sequencing reads (or, correspondingly, the proportion of RNA in a specified sample) contributed by the mother. Subtracting the maternal contribution from one gives the fetal contribution, or the proportion of sequencing reads contributed by the fetus. At step 1511, fetal contribution is calculated for the selected informative maternal loci.

At step 1512, the portion of RNA originating from the fetus is determined from the sample. The portion indicated represents the mean of the fetal contributions for the selected informative maternal loci. In some embodiments, said average is weighted, for example, by sums calculated for selected informative maternal loci. When calculating the weighted average, the portion determined at step 1512 may reflect the relative expression levels of different loci or genes in the sample.

The portion of fetal-derived RNA in a sample can be calculated for one selected maternal informative locus, many such loci, or all such loci for which sequencing data is obtained. It should be borne in mind that the indicated part reflects the circulating transcriptome in a specific subject - a pregnant woman at the point in time when the indicated one was obtained

- 11 046998 sample. In the event that samples are obtained from a subject at different time points during pregnancy, the set of selected maternal informative loci identified in accordance with method 1500 may vary between samples, as well as the portion of fetal-derived RNA identified at these loci. The portion of fetal-derived RNA in a sample may also differ between subjects, even when controlling for gestational age or other factors, and may differ between a subject having a pregnancy-related disorder and a healthy subject. Accordingly, said portion may be compared with a threshold value for diagnosing such a disorder.

In some embodiments, the threshold value is determined by performing method 1500 using samples from one or more healthy subjects. In some embodiments, a diagnosis is made if the portion is greater than a threshold value or less than a threshold value by any amount or by an amount greater than a certain tolerance. The comparison may include calculating the difference or calculating the ratio between the fetal-derived portion of the RNA in the pregnant female subject and the threshold value. In some embodiments, the comparison includes assessing whether the portion of fetal-derived RNA in a pregnant female subject is significantly different from that observed in healthy pregnant women at similar gestational periods.

The portion of fetal-derived RNA in a sample can also be determined by examining fetal informative loci. At each fetal informative locus, an estimate of the specified portion can be obtained by first calculating the ratio for the fetus, which is the number of sequencing reads containing the second (fetal-specific) allele divided by the total number of sequencing reads. In this case, the fetal contribution is equal to the fetal ratio multiplied by a scalar reflecting the relative expression levels of the specified two alleles in the fetus.

Accordingly, a variant of method 1500 may be performed by identifying fetal informative loci, selecting said loci, calculating the fetal contribution for each selected locus, and averaging the fetal contribution for the selected loci. If necessary, the fetal-derived portion of the RNA determined by method 1500 can be expanded by averaging the fetal contributions calculated for the selected maternal informative loci and the fetal contributions calculated for the selected fetal informative loci. The average values calculated in this manner can be weighted to account for variations in the number of sequencing reads at different loci, to increase the weight of specific maternal or specific fetal loci, etc. if necessary.

B. Examples

1. Identification and assessment of fetal and maternal derived transcript in maternal plasma Informative genes, defined as genes containing at least one informative SNP, were first identified based on genotyping data. In the two early pregnancies, a total of 6714 and 6753 informative genes were available to examine the relative proportions of fetal and maternal contributions, respectively (Figs. 16 and 17). In the two late pregnancy cases, a total of 7788 and 7761 informative genes were available to examine the relative proportions of fetal and maternal contributions, respectively. To measure the relative proportion of fetal contribution in maternal plasma, we selected RNA transcripts that contained fetal-specific alleles covered by at least one RNA-seq read in maternal plasma samples. The relative abundance of these fetal-derived transcripts was 3.70 and 11.28% in early and late gestation, respectively. Using a similar approach, the relative proportions of maternal contribution to the blood flow examined using specific maternal SNP alleles were found to be 76.90 and 78.32% in early and late gestation, respectively (Fig. 18A).

2. Comparison of relative fetal and maternal contributions to RNA-SNPs for PET and control pregnancies

The authors examined the relative contributions of the fetus and mother to the GNAS transcript, detected in both cases, both PET, 5641, and in the control case corresponding to its gestational age, 7171. The GNAS transcript contained an informative SNP site containing a fetus-specific SNP allele in locus rs7121 in case 5641; at the same SNP site in case 7171, a maternal-specific SNP allele was present. The calculated relative contributions of the fetus and mother for the specified SNP site in case 5641 were 0.09 and 0.91, respectively. On the other hand, the fetal and maternal specific contributions for the same SNP site in case 7171 were 0.08 and 0.92, respectively (Fig. 19). Compared to control case 7171, there was a 12.5% increase in relative fetal contribution and a 1.09% decrease in relative maternal contribution to the indicated transcript. Interestingly, in the case of 5641, an increase in FPKM of 21% relative to 7171 was detected for the indicated transcript.

- 12 046998

IV. Method for determining a genomic locus as a maternal or fetal marker

A. Method

A method 2000 for determining that a genomic locus is a maternal or fetal marker is presented in FIG. 20. This method involves analyzing a sample obtained from a subject - a woman who is pregnant with the fetus. The sample may be a maternal plasma sample and contains a mixture of RNA molecules of maternal or fetal origin. The sample may be obtained and prepared as discussed above for method 300.

At step 2001, a set of sequencing reads is obtained. At step 2002, said sequencing reads are aligned to a reference sequence. These steps may be performed as described above for steps 301 and 302.

At step 2003, one or more informative loci are identified. Each locus is homozygous in the first object for the corresponding first allele and heterozygous in the second object for the corresponding first allele and the corresponding second allele. The first object represents the subject of the pregnant woman or the fetus, and the second object represents another object of the subject of the pregnant woman and the fetus. Informative loci may be identified as described above at step 303.

At step 2004, information loci are selected in the same manner as in step 304, and one or more selected information loci are identified. In some embodiments, only informative loci representing SNPs are considered for selection. According to one example implementation of this method, the selected informative loci included informative SNPs containing at least one sequence read for the A allele and one sequence read for the B allele in maternal plasma.

Sequencing reads are then processed at the level of individual informative loci selected at stages 2005-2008. In steps 2005 and 2006, a first number and a second number are determined for each selected information locus. Said first number is determined in step 2005 as the number of sequencing reads aligned to said locus and containing the corresponding first allele. Said second number is determined in step 2006 as the number of sequencing reads aligned to said locus and containing the corresponding second allele. Steps 2005 and 2006 may be performed similarly to steps 305 and 306 described above.

At step 2007, the ratio of said first number and said second number is calculated for each selected information locus. This ratio reflects the relative abundance of sequencing reads (and thus transcripts in a sample) containing the first allele and the second allele. In some embodiments, the ratio is simply calculated as the first number divided by the second number. This calculation gives the number of sequencing reads for a common allele as a set of sequencing reads for specific maternal or specific fetal alleles. This ratio, or its inverse, can be considered an allelic ratio, and for SNP loci, an RNA-SNP allelic ratio.

In some embodiments, for informative SNPs that include maternal-specific alleles of the SNP, the RNA-SNP allelic ratio for each SNP is calculated as the maternal-specific allele:total allele ratio. On the other hand, for informative SNPs that include fetal-specific SNP alleles, the RNA-SNP allelic ratio can be calculated as the fetal-specific allele:total allele ratio. Unlike gene expression analysis, in which data normalization is a prerequisite, analysis of RNA-SNP allelic relationships does not require data normalization and introduces less systematic error. For transcripts containing more than one informative SNP, the RNA-SNP allelic ratio for each informative SNP site can be calculated and the average RNA-SNP allelic ratio per transcript can be calculated.

At step 2007, the ratio may alternatively be calculated as the second number divided by the sum of said first number and said second number. In this case, the specified ratio represents the number of sequencing reads for the specific maternal or fetal specific allele as a proportion of all sequencing reads at the specified locus. In the case of alleles A and B, this ratio can be expressed as

Ratio for allele B = counted amount of allele B / (counted amount of allele A + counted amount of allele B)

Theoretically, for a plasma transcript contributed exclusively by the fetus or mother, the ratio for the B allele should be 0.5, assuming allele-specific expression. Other methods for calculating this ratio will be apparent to one skilled in the art.

At step 2008, the selected informative locus is determined to be a marker if the specified ratio exceeds a threshold value. In some embodiments, the threshold value is from about 0.2 to about 0.5. In some embodiments, the threshold value is 0.4. According to one of the examples of implementation of this method, it is possible

- 13 046998 The value for the B allele ratio was defined as >0.4 for an RNA transcript with high fetal or maternal contribution. The specified threshold took into account the Poisson distribution and random selection of RNA-seq reads. In some embodiments, the contribution of the B allele is said to be high if the ratio exceeds a threshold value.

A high allelic ratio at a particular informative locus may indicate that (i) the level of expression of the second allele, or allele B, in a heterozygous individual is higher than the level of expression of allele A (i.e. the allele is expressed asymmetrically), (ii) a greater the proportion of all RNAs aligned to the specified locus in maternal plasma is contributed by the heterozygous individual; or both (i) and (ii) are true. A high allelic ratio may also indicate a pregnancy-related disorder if the gene associated with the indicated informative locus is pathologically overexpressed. Accordingly, some embodiments of the method also include diagnosing a pregnancy-related disorder based on whether the selected informative locus is identified as a marker of a second entity (ie, a heterozygous individual). When making such a diagnosis, the cutoff value used for comparison with allelic ratio may be based on transcription levels in plasma samples obtained from healthy pregnant subjects.

In some embodiments, the method 2000 may also be used to evaluate a portion of the maternal or fetal RNA in a sample at a particular message locus selected. Said estimate is made by multiplying said ratio calculated at step 2007 by a scalar. The specified scalar reflects the overall level of expression at the specified locus relative to the expression of the second allele in a heterozygous individual.

For example, if the second (B) allele counted in the method is fetal specific, then the B allele ratio calculated as described above reflects the number of sequencing reads for the specified allele as a proportion of all sequencing reads for the selected informative locus. Of the sequencing reads containing the common (A) allele, some were contributed by the mother and some were contributed by the fetus. If the relative expression levels of fetal specific and fetal common alleles are known or can be estimated, the B allele ratio can be scaled to provide an estimate of the relative fetal contribution to all sequencing reads for a given locus. If alleles A and B are expressed equally in the fetus, the scalar is approximately 2. In this case, subtracting the relative fetal contribution from one gives the fractional maternal contribution to the sequencing reads at the specified locus. Similar calculations can be performed if the B allele in the selected informative locus is maternal specific.

B. Example: high fetal and maternal contributions

Using a cutoff value of 0.4 for the B allele as described above, high fetal contribution in early gestation (i.e., first and second trimesters) was found to be observed for 0.91% of circulating transcripts. This percentage increased to 2.52% in late gestation (i.e., third trimester). Moreover, a high maternal contribution was observed for 42.58 and 50.98% of transcripts in early and late gestation, respectively (Figs. 16, 17, and 18A).

V. Use of genes to diagnose pregnancy-related disorders

A. Pregnancy-related genes

A method for identifying pregnancy-related genes is also proposed. The method includes obtaining a set of first sequencing reads and a set of second sequencing reads. These first sequencing reads are obtained by sequencing RNA molecules obtained from a blood plasma sample of a pregnant woman. These second sequencing reads are obtained by sequencing RNA molecules obtained from a blood plasma sample of a non-pregnant woman. These first sequencing reads and second sequencing reads are aligned to a reference sequence, and a set of candidate genes is determined.

In this method, the sequencing reads are then used to determine expression levels for each candidate gene in samples from a pregnant woman and a non-pregnant woman. Specifically, for each candidate gene, the first number of transcripts corresponding to said candidate gene is determined using the first sequencing reads; and a second number of transcripts corresponding to said candidate gene is determined using the second sequencing reads. Said first number of transcripts and said second number of transcripts may be normalized. The transcript ratio for the candidate gene is then calculated, wherein the transcript ratio comprises the first number of transcripts divided by the second number of transcripts. The transcript ratio is then compared to a threshold value. The candidate gene is identified as a pregnancy-associated gene if the transcript ratio exceeds a threshold value.

According to some embodiments of this method, normalizing the first number of transcripts corresponds to scaling the first number of transcripts by the total number of first sequencing reads. Similarly, normalizing the second number of transcripts can correspond

- 14 046998 scale the second number of transcripts by the total number of second sequencing reads. In other embodiments, normalizing the first number of transcripts for each candidate gene corresponds to scaling the first number of transcripts for the candidate gene by the total number of first transcripts for all candidate genes. Normalizing the second number of transcripts for each candidate gene may correspond to scaling the second number of transcripts for the candidate gene by the total number of second transcripts for all candidate genes.

In this method, genes are generally identified as being associated with pregnancy if their expression levels are higher in a pregnant woman compared to a non-pregnant woman, other things being equal. The pregnant and non-pregnant woman may be the same individual; those. samples obtained from an individual before and after birth may be the source of first sequencing reads and second sequencing reads, respectively.

The authors showed that a subset of circulating RNA transcripts carrying specific fetal alleles completely disappeared from maternal plasma after delivery (Fig. 18B). These transcripts were considered fetal specific in maternal plasma. On the other hand, some specific maternal alleles were also not detected after childbirth. Accordingly, the authors examined genes, called pregnancy-associated genes, that were up-regulated during pregnancy by directly comparing their representation in maternal plasma before and after delivery. The authors defined pregnancy-associated genes as genes that were detected in the plasma of pregnant women before delivery in the third trimester and whose levels were reduced >2-fold in postpartum plasma in both cases. Using a bioinformatics algorithm to normalize data and analyze differential gene expression, the authors compiled a list of 131 pregnancy-associated genes (Fig. 21). Of these genes, 15 were previously described as pregnancy-specific in maternal plasma. Using one-step real-time RT-PCR, the authors further confirmed the association with pregnancy for five newly identified transcripts that were present at high levels in maternal plasma before delivery, namely STAT1, GBP1 and HSD17B1 in 10 additional plasma samples from pregnant women in the third trimester. , as well as KRT18 and GADD45G in 10 plasma samples from another cohort of pregnant women in the third trimester (Fig. 22).

To assess the association of these 131 genes with pregnancy, hierarchical clustering was performed for all plasma samples. There was a clear separation of plasma samples from pregnant women (namely, early and late pregnancy) and samples not associated with the current pregnancy (namely, from non-pregnant controls and postpartum) (Fig. 23).

Interestingly, when comparing the expression patterns of these 131 genes in plasma samples from two late pregnancies and their corresponding placentas and maternal blood cells, closer similarities were observed between placenta and plasma samples before delivery, and between maternal blood cells and plasma samples after delivery. childbirth (Fig. 24A). This observation supports the idea that most pregnancy-related genes are preferentially expressed in the placenta rather than in maternal blood cells. In addition, the expression levels of these pregnancy-related transcripts in the placenta and maternal plasma were positively correlated (P < 0.05, Spearman correlation) (Fig. 24B).

B. Differential gene expression in the placenta and maternal blood

While the authors were able to catalog a panel of pregnancy-related genes by directly examining antepartum and postpartum maternal plasma samples, the authors also analyzed placental and blood cell RNA-seq data for comparison purposes. Under the assumption that pregnancy-associated genes would be represented by genes with high levels of expression in the placenta and low levels of expression in maternal blood cells, as previously reported, the authors established a conditional 20-fold difference level as the minimum threshold in the tissue-based analysis. This tissue-based analysis yielded a total of 798 potential candidate genes for which the proportions of fetal and maternal contributions to maternal plasma were calculated. A relatively high proportion of genes with predominant fetal contributions were identified (Figure 25) compared to the full transcriptome (Figure 18A). However, the plasma-based strategy was superior to this tissue-based strategy, allowing the identification of a larger proportion of genes with a predominance of fetal contribution (Fig. 25).

C. Genes associated with disease or disorder

The present invention also provides a method for identifying genes associated with a pregnancy-related disorder. The method includes obtaining a set of first sequencing reads and a set of second sequencing reads. These first sequencing reads are obtained by sequencing RNA molecules obtained from a blood plasma sample of a healthy pregnant woman. These second sequencing reads are obtained by sequencing RNA molecules obtained from a blood plasma sample of a pregnant woman suffering from a pregnancy-related disorder or carrying a fetus suffering from a pregnancy-related disorder. Specified

- 15 046998 The specified first sequencing reads and the specified second sequencing reads are aligned with the reference sequence, and a set of candidate genes is determined.

In this method, the sequencing reads are then used to determine expression levels for each candidate gene in samples from two pregnant women. Specifically, for each candidate gene, a first transcript number corresponding to said candidate gene is determined using first sequencing reads, and a second transcript number corresponding to said candidate gene is determined using second sequencing reads. Said first number of transcripts and said second number of transcripts may be normalized. The transcript ratio for the candidate gene is then calculated, wherein the transcript ratio comprises the first number of transcripts divided by the second number of transcripts. The reported transcript ratio is then compared to a reference value. A candidate gene is identified as being associated with a disorder if the reported transcript ratio deviates from the reference value.

According to some embodiments of this method, normalizing the first number of transcripts corresponds to scaling the first number of transcripts by the total number of first sequencing reads. Likewise, normalizing the second number of transcripts may correspond to scaling the second number of transcripts by the total number of second sequencing reads. In other embodiments, normalizing the first number of transcripts for each candidate gene corresponds to scaling said first number of transcripts for the candidate gene by the total number of first transcripts for all candidate genes. Normalizing the second number of transcripts for each candidate gene may correspond to scaling said second number of transcripts for the candidate gene by the total number of second transcripts for all candidate genes.

According to a method for identifying genes associated with a pregnancy-associated disorder, in some embodiments, the reference value is 1. In some embodiments, said transcript ratio deviates from the reference value if the ratio of said transcript ratio and the reference value is above a threshold value or below a threshold value . In some embodiments, the specified transcript ratio deviates from the reference value if the difference between the transcript ratio and the reference value exceeds a threshold value.

In this method, typically a gene is identified as associated with a pregnancy-related disorder if the level of expression of the gene is significantly different between pregnancies with and without the disorder, all other things being equal.

Diagnosis and monitoring of fetal disorders and pregnancy pathologies have previously been carried out using maternal plasma RNA associated with diseases. For example, it has been shown that maternal plasma levels of corticotropin-releasing hormone (CRH) mRNA can be used to noninvasively detect and predict preeclampsia. Detection of interleukin-1 receptor-like protein (IL1RL1) mRNA in maternal plasma has been demonstrated to be useful in identifying women predisposed to spontaneous preterm birth. A panel of maternal plasma growth-related RNA markers for non-invasive assessment of fetal growth and intrauterine growth restriction was also examined. In the present study, the authors conclude that novel disease-associated circulating RNA markers can be identified by directly comparing maternal plasma transcriptomes from healthy pregnant women and women with pregnancy complications such as preeclampsia, intrauterine growth restriction, preterm birth, and fetal aneuploidies. To demonstrate the applicability of this approach, the authors performed RNA-Seq on maternal plasma samples obtained from three pregnant women who developed preeclampsia and seven uncomplicated pregnant women at matched gestational ages. The authors identified 98 transcripts that were significantly elevated in the plasma of pregnant women with preeclampsia (Figure 26). For the first time, preeclampsia-associated transcripts have been identified with potential applications for prediction, prognosis and monitoring of the disease.

This technique can also be used to predict and monitor preterm birth. This technique can also be used to predict the risk of fetal death. This technique can also be used to detect diseases caused by gene mutations, provided that the gene involved is transcribed in fetal or placental tissues and the transcripts are detected in maternal plasma.

Plasma RNA-Seq has potential applications in other clinical settings. For example, the plasma RNASeq technology developed in this study could potentially be suitable for the study of other pathological conditions, such as cancer, for which aberrant plasma RNA concentrations have been described. For example, by comparing the plasma transcriptomes of a cancer patient before and after therapy, tumor-associated circulating RNA markers can be identified for use in non-invasive diagnostics.

- 16 046998

VI. Expression patterns of alleles of specific genes

RNA sequencing is used to study the expression pattern of alleles. The authors postulated that the expression pattern of the alleles of a particular gene is maintained when RNA transcripts are released from tissues into the bloodstream and therefore can be detected in plasma. In the present study, the authors examined the counted number of alleles in two RNA transcripts, namely PAPPA, a pregnancy-specific gene, and H19, a maternally expressed imprinting gene.

In the case of the PAPPA gene, we analyzed the SNP, rs386088, which contains a specific fetal allele SNP (Fig. 27). The absence of RNA-seq reads for PAPPA in postpartum plasma samples indicates that it is indeed pregnancy specific. Note that there was no statistically significant difference in the proportions of counted reads for the fetal allele in RNA from antenatal and placental maternal plasma samples (P = 0.320, ^χ2 test), indicating that the maternal plasma data reflect a biallelic pattern of PAPPA expression in placenta.

Recently, the authors described the possibility of determining, by bisulfite DNA sequencing of maternal plasma, the DNA methylation pattern of the imprinted H19 gene expressed from the maternal allele in the placenta and maternal blood cells. In this work, the authors further investigated whether it is possible to study the genomic imprinting status of the H19 gene at the RNA level. The authors first focused on a SNP site in exon 1 of the H19 gene, rs2839698, which contains a maternal-specific allele, namely AA in the fetus and AG in the mother. As can be seen from Fig. 28, only the G allele was detected in postpartum maternal plasma (Fig. 28). This monoallelic pattern was consistent with an association with the unmethylated G allele at SNP site rs4930098 in the imprinting control region (Fig. 29). In maternal plasma before birth, in addition to the presence of the G allele, the presence of the A allele introduced by the placenta was also detected (Fig. 28). Three other SNP sites, namely, rs2839701, rs2839702 and rs3741219, carrying specific maternal alleles, showed a similar allelic pattern, namely, biallelic in prepartum maternal plasma and monoallelic in postpartum maternal plasma (Fig. 28). It is possible that specific maternal alleles at these SNP sites were located in the same maternal haplotype that was not methylated and thus transcribed. Notably, H19 RNA was not expressed in maternal blood cells (Figure 28), suggesting that H19 RNA molecules in plasma originate from other maternal tissues/organs rather than blood cells. Non-placental and non-fetal tissues that have been reported to express H19 include adrenal gland, skeletal muscle, uterus, adipocytes, liver and pancreas.

VII. Discussion

The goal of this work was to develop a technology for obtaining a global picture of transcriptomic activity in maternal plasma using RNA-seq (RNA sequencing). The authors previously showed that the proportion of fetal DNA in maternal plasma can be calculated from one or more target fetal-specific loci, since the entire fetal genome is evenly represented in maternal plasma. In contrast to measuring circulating DNA, measuring the proportion of fetal-derived RNA transcripts in maternal plasma is not as straightforward, as it is complicated by differential gene expression in fetal and maternal tissues and their possible release into the circulation. By performing RNA-seq on maternal plasma and examining fetal-maternal polymorphic differences, the authors were able to estimate the proportion of fetal-contributed transcripts in plasma. Although the plasma transcriptome was predictably dominated by maternally derived transcripts, 3.70 and 11.28% of circulating transcripts in maternal plasma in early and late pregnancy, respectively, were of fetal origin. These fetal-derived transcripts include RNA molecules contributed jointly by the fetus and mother, as well as those contributed solely by the fetus. The authors found the latter value to be 0.90 and 2.52% of maternal circulating transcripts in early and late pregnancy, respectively. The greater representation of such specific fetal genes in late pregnancy may be due to an increase in the size of the fetus and placenta as pregnancy progresses.

In the present study, the authors demonstrated that in maternal plasma there is a balanced allelic RNA expression of the pregnancy-specific PAPPA gene in the placenta and monoallelic expression of the imprinted H19 gene expressed from the maternal allele. These data suggest that maternal plasma can be used as a source of samples for the non-invasive study of allelic expression patterns.

Based on a quantitative comparison of RNA transcripts in antepartum and postpartum maternal plasma samples, the authors compiled a list of 131 genes that are up-regulated during pregnancy, which is evident from the decreased representation of these genes in postpartum plasma samples. As expected, the profile of these genes can be used to differentiate plasma samples from pregnant women from plasma samples from non-pregnant women. This

- 17 046998 direct comparison of prepartum and postpartum maternal plasma samples allowed the authors to select, with high throughput, pregnancy-associated genes whose expression levels in the placenta are not necessarily significantly higher than in maternal blood cells, as demonstrated in a previous study. In essence, this direct plasma assay represents a new approach to detect circulating pregnancy-associated RNA transcripts in the absence of a priori information about the transcriptomic profiles of placental tissues and blood cells.

Although the authors showed that RNA-seq is a suitable method for profiling the plasma transcriptome, some technical aspects could be further improved. First, the amount of information obtained from plasma RNA-seq can be increased by further optimization of the sequencing protocol, in particular with regard to plasma depletion of highly transcribed rRNA and globin genes. Second, the authors focused only on reference transcripts, and individual isoforms have not yet been examined. Future studies may include detection of new transcripts and differential analysis of splice variants and their isoforms by increasing the read depth of sequencing. Third, the authors excluded selection by ASE when analyzing the proportions of fetal and maternal transcripts for samples obtained in early pregnancy, since chorionic villi and amniotic fluid were consumed in the genotypic analysis. However, the authors showed that in late pregnancy ASE had no significant effect on the identification of genes with predominant fetal and maternal contributions in maternal plasma (Figs. 16 and 17).

Finally, the authors demonstrated that RNA-seq technology can be used to measure the proportion of fetal-derived transcripts and to identify circulating pregnancy-associated genes in maternal plasma. The present study sets the stage for a better understanding of the transcriptomic landscape of maternal plasma, facilitating the identification of candidate biomarkers involved in pregnancy-related diseases. The authors believe that this technology will lead to new approaches to molecular diagnostics of pregnancy- and placenta-related diseases, as well as other diseases such as cancer.

VIII. Transplantation

As mentioned above, the methods described herein can also be used in transplantation. Methods for use in transplantation can be carried out in the same manner as in fetal research. For example, the genotype of the transplanted tissue can be obtained. Embodiments allow for the identification of loci where the transplanted tissue is heterozygous and the host organism (eg, male or female) is homozygous. Also, embodiments make it possible to identify loci where the transplanted tissue is homozygous and the host organism (for example, male or female) is heterozygous. The same ratios can be calculated and compared to a threshold to determine whether a violation exists.

IX. Computer systems

Any suitable number of subsystems may be used in any of the computer systems referred to herein. Examples of such subsystems are presented in Fig. 30 in computing device 3000. In some embodiments, a computer system includes a single computing device, wherein the subsystems may be components of said computing device. In other embodiments, a computer system may include a plurality of computer devices, each of which is a subsystem, with internal components.

The subsystems shown in FIG. 30 are connected by system bus 3075. Additional subsystems are shown, such as a printer 3074, a keyboard 3078, a storage device(s) 3079, a monitor 3076 that is connected to a display adapter 3082, and others. Peripherals and input/output (I/O) devices associated with the I/O controller 3071 may be connected to the computer system by any means known in the art, such as through a serial port 3077. Thus, a serial port 3077 or an external interface 3081 (eg, Ethernet, Wi-Fi, etc.) may be used to connect computer system 3000 to a wide-area network such as the Internet, a mouse input device, or a scanner. The connection through the system bus 3075 allows the central processing unit 3073 to communicate with each subsystem and control the execution of instructions from the system memory 3072 or storage device(s) 3079 (e.g., a permanent disk such as a hard drive or optical disk), as well as the exchange of information between the subsystems. System memory 3072 and/or storage device(s) 3079 may be computer-readable media. Any data referred to herein may be output from one component to another component, and may be output to be provided to the user.

The computer system may include a collection of the same components or subsystems, for example, connected by an external interface 3081 or an internal interface. In some embodiments, the computer systems, subsystem, or hardware may communicate with each other via a network. In such cases, one computer can be considered the client, and the other computer

- 18 046998 as a server, and each of them can 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 may be implemented in the form of a logic control device using hardware (eg, an application specific integrated circuit or a programmable gate array) and/or using computer software using a conventional programmable processor in a modular or integrated manner. As used herein, a processor includes a multi-core processor on a single integrated circuit or a plurality of processing modules on a single board circuit or network structure. Based on the present disclosure and the principles set forth herein, other approaches and/or methods for implementing embodiments of the present invention using hardware and combinations of hardware and software will be apparent to one of ordinary skill in the art.

Any software components or functions described in this application may be implemented as program code for execution by a processor using any suitable computer language, such as, for example, Java, C++ or Perl, using, for example, traditional or object-oriented technician. The program code may be stored as a sequence of instructions or commands on a computer readable medium for storing and/or transmitting data, suitable media including random access memory (RAM), read only memory (ROM), magnetic medium such as a hard disk or floppy disk, or an optical media such as a compact disc (CD) or DVD (digital versatile disk), flash memory, etc. A computer-readable medium may be any combination of such data 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 various protocols, including the Internet. Thus, a computer-readable medium according to an embodiment of the present invention can be obtained using a data signal encoded by such programs. The computer-readable medium containing the program code may be packaged with a compatible device or provided separately from other devices (for example, by downloading over the Internet). Any such machine-readable media may be located on or included in a single computer product (e.g., a hard disk drive, CD-ROM, or entire computer system), and may be located on or included in different computer products within a system or networks. The computer system may include a monitor, printer, or other suitable display device for providing the user with any of the results referred to herein.

Any of the methods described herein may be implemented in whole or in part by a computer system including one or more processors that may be configured to carry out the steps. Accordingly, embodiments may relate to computer systems configured to implement steps of any of the methods described herein, potentially with different components implementing the respective steps or a corresponding group of steps. Although the steps of the methods presented herein are numbered in the description, they may be performed simultaneously or in a different order. In addition, parts of these steps may be used with parts of other steps related to other methods. Also, a stage may be completely or partially optional. In addition, any steps of any methods may be implemented using modules, circuits, or other methods of performing said steps.

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

The above description of embodiments of the present invention is provided for illustrative and descriptive purposes. It is not intended to be exhaustive or limiting of the present invention to the precise form described, and many modifications and variations are possible in light of the foregoing principles. The embodiments have been selected and described to best explain the principles of the present invention and its practical application in order to enable others skilled in the art to best use the present invention according to various embodiments and with various modifications suitable for the particular intended application.

All terms used in the singular, including those supplemented by the definition specified, are intended to mean one or more unless specifically stated otherwise.

All patents, patent applications, publications and specifications referred to herein are

- 19 046998 are incorporated by reference in their entirety and for all purposes. None of them are considered prior art.

X. References

1. Ng EKO, Tsui NBY, Lau TK, Leung TN, Chiu RWK, Panesar NS, Lit LC, Chan KW, Lo YMD. mRNA of placental origin is readily detectable in maternal plasma. Proc Natl Acad Sci US A2003;100:4748-53.

2. Ng EKO, Leung TN, Tsui NBY, Lau TK, Panesar NS, Chiu RWK, Lo YMD. The concentration of circulating corticotropin-releasing hormone mRNA in maternal plasma is increased in preeclampsia. Clin Chern 2003;49:727-31.

3. Farina A, Sekizawa A, Sugito Y, Iwasaki M, Jimbo M, Saito H, Okai T. Fetal DNA in maternal plasma as a screening variable for preeclampsia. A preliminary nonparametric analysis of detection rate in low-risk nonsymptomatic patients. Prenat Diagn 2004;24:83-6.

4. Pang WW, Tsui MH, Sahota D, Leung TY, Lau TK, Lo YM, Chiu RW. A strategy for identifying circulating placental RNA markers for fetal growth assessment. Prenat Diagn 2009;29:495-504.

5. Lo YMD, Tsui NBY, Chiu RWK, Lau TK, Leung TN, Heung MM, Gerovassili A, Jin Y, Nicolaides KH, Cantor CR, Ding C. Plasma placental RNA allelic ratio permits noninvasive prenatal chromosomal aneuploidy detection. Nat Med 2007;13:218-23.

6. Tsui NBY, Akolekar R, Chiu RWK, Chow KCK, Leung TY, Lau TK, Nicolaides KH, Lo YMD. Synergy of total PLAC4 RNA concentration and measurement of the RNA singlenucleotide polymorphism allelic ratio for the noninvasive prenatal detection of trisomy 21. Clin Chern 2010;56: 73-81.

7. Tsui NBY, Wong BCK, Leung TY, Lau TK, Chiu RWK, Lo YMD. Non-invasive prenatal detection of fetal trisomy 18 by RNA-SNP allelic ratio analysis using maternal plasma SERPINB2 mRNA: a feasibility study. Prenat Diagn 2009;29:1031-7.

8. Chim SS, Lee WS, Ting YH, Chan OK, Lee SW, Leung TY. Systematic identification of spontaneous preterm birth-associated RNA transcripts in maternal plasma. PLoS One 2012;7:e34328.

9. Wong BCK, Chiu RWK, Tsui NBY, Chan KS A, Chan LW, Lau TK, Leung TN, Lo YMD. Circulating placental RNA in maternal plasma is associated with a preponderance of 5' mRNA fragments: implications for noninvasive prenatal diagnosis and monitoring. Clin Chern 2005;51:1786-95.

10. Tsui NBY, Chim SSC, Chiu RWK, Lau TK, Ng EKO, Leung TN, Tong YK, Chan KCA, Lo YMD. Systematic micro-array based identification of placental mRNA in maternal plasma: towards non-invasive prenatal gene expression profiling. J Med Genet 2004;41:461-7.

11. Poon LL, Leung TN, Lau TK, Lo YMD. Presence of fetal RNA in maternal plasma. Clin Chern 2000;46:1832-4.

- 20 046998

12. Go AT, Visser A, Mulders MA, Blankenstein MA, Van Vugt JM, Oudejans CB. Detection of placental transcription factor mRNA in maternal plasma. Clin Chern 2004;50:14134.

13. Smets EM, Visser A, Go AT, van Vugt JM, Oudejans CB. Novel biomarkers in preeclampsia. Clin Chim Acta 2006;364:22-32.

14. Purwosunu Y, Sekizawa A, Koide K, Farina A, Wibowo N, Wiknjosastro GH, et al. Cell-free mRNA concentrations of plasminogen activator inhibitor-1 and tissue-type plasminogen activator are increased in the plasma of pregnant women with preeclampsia. Clin Chern 2007;53:399-404.

15. Miura K, Miura S, Yamasaki K, Shimada T, Kinoshita A, Niikawa N, et al. The possibility of microarray-based analysis using cell-free placental mRNA in maternal plasma. Prenat Diagn 2010;30:849-61.

16. Ng EKO, El-Sheikhah A, Chiu RWK, Chan KS, Hogg M, Bindra R, et al. Evaluation of human chorionic gonadotropin beta-subunit mRNA concentrations in maternal serum in aneuploid pregnancies: a feasibility study. Clin Chern 2004;50:1055-7.

17. Mortazavi A, Williams BA, McCue K, Schaeffer L, Wold B. Mapping and quantifying mammalian transcriptomes by RNA-Seq. Nat Methods 2008;5:621-8.

18. Sultan M, Schulz MH, Richard H, Magen A, Klingenhoff A, Scherf M, et al. A global view of gene activity and alternative splicing by deep sequencing of the human transcriptome. Science 2008;321:956-60.

19. Kim J, Zhao K, Jiang P, Lu ZX, Wang J, Murray JC, Xing Y. Transcriptome landscape of the human placenta. BMC Genomics 2012;13:115.

20. Wang K, Li H, Yuan Y, Etheridge A, Zhou Y, Huang D, et al. The complex exogenous RNA spectra in human plasma: an interface with human gut biota? PLoS One 2012;7:e51009.

21. Li H, Guo L, Wu Q, Lu J, Ge Q, Lu Z. A comprehensive survey of maternal plasma miRNAs expression profiles using high-throughput sequencing. Clin Chim Acta 2012;413:56876.

22. Williams Z, Ben-Dov IZ, Elias R, Mihailovic A, Brown M, Rosenwaks Z, Tuschl T. Comprehensive profiling of circulating microRNA via small RNA sequencing of cDNA libraries reveals biomarker potential and limitations. Proc Natl Acad Sci U S A 2013; 110:4255-60.

23. Chim SSC, Shing TK, Hung EC, Leung TY, Lau TK, Chiu RWK, Lo YMD. Detection and characterization of placental microRNAs in maternal plasma. Clin Chern 2008;54:482-90.

24. Pickrell JK et al. Understanding mechanisms underlying human gene expression

- 21 046998 variation with RNA sequencing. Nature 2010;464:768-772.

25. Smith RM, Webb A, Papp AC, Newman LC, Handelman SK, Suhy A, et al. Whole transcriptome RNA-Seq allelic expression in human brain. BMC Genomics 2013; 14:571.

26. Frost JM, Monk D, Stojilkovic-Mikic T, Woodfine K, Chitty LS, Murrell A, et al.

Evaluation from allex expression of imprinted genes in adult human blood. PLoS One 2010;5:el3556.

27. Daelemans C, Ritchie ME, Smits G, Abu-Amero S, Sudbery IM, Forrest MS, et al. High-throughput analysis of candidate imprinted genes and allele-specific gene expression in the human term placenta. BMC Genet 2010; 11:25.

28. Lun FMF, Chiu RWK, Sun K, Leung TY, Jiang P, Chan KS A, et al. Noninvasive prenatal methylomic analysis by genomewide bisulfite sequencing of maternal plasma DNA. Clin Chern 2013;59:1583-94.

29. Wu C, Orozco C, Boyer J, Leglise M, Goodale J, Batalov S, et al. BioGPS: an extensible and customizable portal for querying and organizing gene annotation resources. Genome Biol 2009;10:R130.

30. Lo YMD, Chan KCA, Sun H, Chen EZ, Jiang P, Lun FMF, et al. Maternal plasma DNA sequencing reveals the genome-wide genetic and mutational profile of the fetus. Sci Transl Med 2010;2:61ra91.

31. Cabili MN, Trapnell C, Goff L, Koziol M, Tazon-Vega B, Regev A, Rinn JL.

Integrative annotation of human large intergenic noncoding RNAs reveals global properties and specific subclasses. Genes Dev 2011;25:1915-27.

32. St Laurent G, Shtokalo D, Tackett MR, Yang Z, Eremina T, Wahlestedt C, et al. Intronic RNAs constitute the major proportion of the non-coding RNA in mammalian cells. BMC Genomics 2012;13:504.

33. Anders S, Reyes A, Huber W. Detecting differential usage of exons from RNA-seq data. Genome Res 2012;22:2008-17.

34. Fleischhacker M, Schmidt B. Circulating nucleic acids (CNAs) and cancer—a survey.

Biochim Biophys Acta 2007;1775:181-232.

Claims (12)

1. A method for diagnosing a pregnancy-related disorder in a female subject who is pregnant with a fetus, the method comprising comparing the expression level of a gene in a biological sample with a control value determined from one or more female subjects, each of the subjects being pregnant with a healthy fetus, wherein said gene represents are ASAH1, C12orf76, C19orf59, C19orf79, C21orf7, C7orf53, CEBPD, DDX11L10, DUSP1, ENKUR, GALM, HIST1H2AC, HIST1H3H, HIST1H4A, HIST1H4B, HIST1H4H, HIST1H4L, HLA-L, IFI6, ISG20, MEIS1, MLH3, MYADM , NCF1B, NES, NGFRAP1, OST4, PF4, PGRMC1, PRIC285, RABAC1, RPPH1, S100P, S1PR3, SIGLEC1, SIGLEC14, SPARC, TMEM140, TMEM185A, TMEM50A, TSC22D3, UBE2L6, or VTRNA1-1; wherein said pregnancy-related disorder is preeclampsia, and the level of gene expression is measured by obtaining a biological sample from a female subject, obtaining multiple sequence reads, wherein the sequence reads are obtained by analyzing cell-free RNA molecules obtained from the biological sample, identifying using a computer system positions of sequence reads in the reference sequence and counting sequence reads located in the gene. 2. The method of claim 1, further comprising diagnosing a pregnancy-related disorder if the expression level exceeds a control value. 3. The method according to claim 1 or 2, wherein the biological sample is a sample of blood, plasma, serum, urine or saliva. 4. The method according to claim 1 or 2, wherein the biological sample is a plasma sample. 5. The method according to claim 1 or 2, wherein the biological sample is a serum sample. 6. Method according to any one of claims 1 to 5, where the genome is ASAH1. 7. Method according to any one of claims 1 to 5, where the gene is TSC22D3. - 22 046998 8. Method according to any one of claims 1 to 5, wherein the gene is MYADM. 9. Method according to any one of claims 1 to 5, where the gene is C7orf53. 10. A system comprising one or more processors configured to implement any of the methods of claims 1 to 9. 11. The system according to claim 10, further comprising a sequencing device. 12. A computer product containing a computer-readable medium storing a set of instructions for controlling a processor to implement the method according to any one of claims 1 to 9.