BE1023266B1 - System and methodology for the analysis of genomic data obtained from a subject - Google Patents

Abstract

The present invention describes a method for determining the presence or absence of a fetal chromosomal aneuploidy in a pregnant woman, the method comprising calculating a parameter p from sequences obtained from a biological sample from said pregnant woman. The present invention also provides a method for determining the fetal fraction of said sample.

Description

SYSTEM AND METHODOLOGY FOR THE ANALYSIS OF GENOMIC DATA OBTAINED FROM A SUBJECT

TECHNICAL FIELD

The invention relates to a method and system for the analysis of genomic data of a patient. The present invention particularly relates to a method that permits the determination of the presence of a fetal aneuploidy in a sample in a substantially non-invasive manner.

PREFACE

The availability of DNA processing technologies with a high processing capacity has allowed almost comprehensive investigations into the number and type of sequence variants in individuals in different populations and with different diseases. Sequencing of the entire genome can become a relatively routine procedure in the near future as the costs and efficiency of sequencing with a high processing capacity continue to improve. Indeed, as costs continue to fall, high-processing sequencing is expected to become a commonly used tool, not only in human phenotype-based sequencing projects, but also as an effective tool in advanced genetic applications in model organisms, and for disease diagnosis previously considered to be idiopathic.

Once a sequence is obtained, an effort is made to identify the location and character of these parts of a sequence that differ from one or more "standard" reference sequences, each of which is commonly referred to as a variant. This can help identify the parts of an individual's genome that could potentially contribute to a clinical condition or other characteristic of the individual. For example, it is common to compare the sequence of a particular individual with human reference genome sequences stored by the University of California, Santa Cruz, and to list the variants that exist between the sequence of an individual and a reference sequence.

This list of variants can contain millions of variants, but gives little or no information about the impact that a particular variant can have on gene function. Research programs worldwide continuously collect information about certain variants of a gene function, disease stages, and the like. In addition, various computer methods have been developed to derive potential physiological effects of certain types of variants based on their location on the genome and the nature of the variant, even if no biochemical or clinical laboratory studies have been performed on that particular variant.

Prenatal analysis of a fetus has today increasingly shifted to a genomic analysis of the fetus and the DNA of the parents.

The presence of circulating extracellular DNA in the peripheral blood is a well-known phenomenon. In the case of a pregnant woman, it has been demonstrated that extracellular fetal DNA is present in the mother's bloodstream and can be detected in the mother's plasma or serum. Although most people in the art describe this type of DNA as fetal DNA, it is in fact placental DNA, and small differences can occur between fetal and placental DNA because of mosaicism that originates during embryogenesis. However, the term "fetal DNA" will be used throughout this document as this is the most commonly used terminology to describe this type of DNA. Studies have shown that this circulating fetal genetic material can be used for the highly reliable determination, e.g. by PCR (polymerase chain reaction) technology, of fetal genetic loci completely absent from the mother's genome. Examples of such fetal genetic loci are the fetal RhD gene during risk pregnancies for HDN (hemolytic disease of the fetus and newborn) or fetal Y chromosome specific sequences in risk pregnancies for an X chromosome-related disorder, e.g. hemophilia or fragile X syndrome.

Fetal aneuploidy and other chromosomal abnormalities affect 9 out of 1,000 live births. The gold standard for diagnosing chromosomal aberrations is karyotyping of fetal cells obtained through invasive procedures such as chorionic villus sampling and amniocentesis. These procedures involve small but potentially significant risks to both the fetus and the mother. In recent years, progress has been made in the development of non-invasive screening methods for fetal chromosomal abnormalities.

Since the discovery of intact fetal cells in the mother's blood, there has been a strong interest in trying to use them as a diagnostic window for fetal genetics. EP 2183693 and family members describe a methodology for performing a prenatal diagnosis of a fetal chromosomal aneuploidy in a biological sample obtained from the pregnant mother. The sample is randomly sequenced and the sequences obtained are used to determine a parameter used to evaluate the presence or absence of fetal aneuploidy. US8195415 describes a method for evaluating whether a chromosome has an abnormal distribution in a sample obtained from a patient. The method includes shotgun sequencing of the DNA present in the sample, and then aligning the resulting sequence tags with chromosome portions. Values are determined based on the number of alignments that are then used to calculate a difference that determines whether or not an abnormal distribution exists.

Bayindir et al., 2015 describes a non-invasive prenatal test methodology based on the Z and ZZ score, where said Z score is a chromosomal broad Z score and the ZZ score is calculated as the standard score of the Z-score of a certain autosome compared to the Z-scores of remaining autosomes.

Although the above methodologies are valuable, the ratio of false positives and negatives remains high in the field. Therefore, there is a constant effort to provide methodologies that can reduce the percentage of false positives, and in particular false negatives, to provide more accurate and robust screening.

The present invention aims to provide a more accurate non-invasive methodology for determining whether there is aneuploidy in a fetus, together with the determination of the fetal fraction as this is considered an important quality metric for assessing the risk of false negatives in a sample.

From a broader perspective, the present invention also aims to provide a methodology and tools for analyzing genomic data, e.g. for genomic variant annotation.

SUMMARY OF THE INVENTION

The invention relates to a method for determining the presence or absence of an aneuploidy according to claim 1 or one of the dependent claims. The methodology according to the claims makes it possible to evaluate the presence or absence of such aneuploidy on the basis of a series of reference samples. The present methodology offers a reliable and robust parameter for determining the presence of aneuploidy. The methodology is sensitive and minimizes false positives and false negatives.

The present invention also provides a method for determining the fetal fraction according to claim 27 and dependent claims. The method provides a clear and easy estimate of the fetal fraction in a sample based on the available random sequencing data with low coverage. The latter also serves as an additional quality control for the detection of aneuploidy, in particular for cases where no aneuploidy has been detected (since some of these could be false positives because the fetal fraction was too low to allow the detection of aneuploidy using sequencing with low coverage).

Finally, the present methodology also provides a computer program product according to claim 19 or 30, which can perform one or more operations according to the present invention and a report generated thereby according to claim 25.

FIGURES

Figure 1 shows a report according to an embodiment of the present invention, showing calculated parameters and a visual representation for chromosome 21 in a sample A, said chromosome 21 being identified as abnormal by the method of the present invention.

Figure 2 shows a graphical representation of the parameters obtained for all chromosomes within one sample A. Only chromosome 21 showed a deviation.

Figure 3 shows a report according to an embodiment of the present invention, showing calculated parameters and a visual representation for chromosome 11 in sample A. No abnormalities have been observed.

Figure 4 shows a graphical representation of the calculated secondary parameters according to an embodiment of the present invention, calculated for all chromosomes of sample A. A trisomy was observed.

Figure 5 shows a report according to an embodiment of the present invention, showing calculated parameters and a visual representation for chromosome 16 in sample B. A trisomy was observed.

Figure 6 shows a graphical representation of the parameters obtained for all chromosomes within sample B. Only one chromosome, chromosome 16, showed a deviation.

Figure 7 shows a report according to an embodiment of the present invention, showing calculated parameters and a visual representation for chromosome 1 in sample B. No trisomy was observed.

Figure 8 shows a graphical representation of the calculated secondary parameters according to an embodiment of the present invention, calculated for all chromosomes of sample B. A trisomy was observed.

Figure 9 shows a report according to an embodiment of the present invention, showing calculated parameters and a visual representation for chromosome 21 in a sample A, said chromosome 21 being identified as abnormal by the method of the present invention.

Figure 10 shows a graphical representation of the parameters obtained for all chromosomes within one sample A. Only one chromosome showed a deviation.

Figure 11 shows a report according to an embodiment of the present invention, showing calculated parameters and a visual representation for chromosome 11 in sample A. No abnormalities have been observed.

Figure 12 shows a report according to an embodiment of the present invention, showing calculated parameters and a visual representation for chromosome 16 in sample B. An aneuploidy was observed.

Figure 13 shows a graphical representation of the parameters obtained for all chromosomes within sample B. Only one chromosome 16 showed a deviation.

Figure 14 shows a report according to an embodiment of the present invention, showing calculated parameters and a visual representation for chromosome 1 in sample B. No trisomy was observed.

Figure 15 shows a graphical representation of a secondary parameter indicative of the success rate of the experiment obtained for all chromosomes in sample A according to an embodiment of the present example. The experiment was successful.

Figure 16 shows a graphical representation of a secondary parameter indicative of the success rate of the experiment obtained for all chromosomes in sample B according to an embodiment of the present example. The experiment was successful.

Figure 17 shows a graphical representation in a sample obtained by an embodiment of the present invention indicating the presence of a polymorphic site, more particularly a copy number variation that may be present in the mother's genome instead of the genome of the fetus.

Figures 18 to 24 show graphical representations for a chromosome in a sample obtained by an embodiment of the present invention. The methodology of the present invention has been proven to have a higher sensitivity than the methods of the prior art.

Figure 25 shows a graphic representation of the number of readings with respect to the Y chromosome for samples classified as male, female, or indeterminate.

Figure 26 shows a graphical representation of the X and Y-based fetal fraction estimates for a series of male pregnancies.

Figure 27 shows the scatter diagrams of the normalized counts for a particular polymorphic site within a series of test samples.

Figures 28 and 29 show a graphical representation of the estimated fetal fraction for a series of male pregnancies based on a particular informative polymorphic site, as calculated using an embodiment of the present invention (X-axis) versus based on the reading counts of chromosome X or Y (Y-axis).

Figure 30 shows a graphic representation of the estimated fetal fraction for a series of male pregnancies based on the informative polymorphic site as identified in the sample as calculated using an embodiment of the present invention (X-axis) versus the estimated fetal fraction based on the reading counts of chromosome X or Y (Y-axis).

DEFINITIONS

The term "biological sample" as used herein refers to any sample taken from a patient (e.g., a human, such as a pregnant woman) and containing one or more interesting nucleic acid molecule (s).

The term "nucleic acid" or "polynucleotide" refers to a deoxyribonucleic acid (DNA) or ribonucleic acid (RNA) and a polymer thereof in either single or double-stranded form. Unless specifically limited, the term includes nucleic acids containing well-known analogs of natural nucleotides that have similar binding properties to the reference nucleic acid and are similarly metabolized as naturally occurring nucleotides. Unless otherwise stated, a particular nucleic acid sequence also includes implicitly conversatively modified variants thereof (e.g., degenerator codon substitutions), alleles, orthologs, single-nucleotide polymorphisms (SNPs) and complementary sequences as well as the explicitly indicated sequence. Degeneration codon substitutions can in particular be obtained by generating sequences in which the third position of one or more selected (or all) codons is substituted with mixed-base and / or deoxyinosine residues. The term nucleic acid is used interchangeably with gene, DNA, cDNA, mRNA, small non-coding RNA, micro-RNA (miRNA), Piwi-interacting RNA, and short hairpin RNA (shRNA) encoded by a gene or locus.

The term "gene" means the segment of DNA that is involved in the production of a polypeptide chain. It may include regions preceding or following the coding region (head and tail) as well as intervening sequences (introns) between individual coding segments (exons).

The term "reaction" as used herein refers to any process comprising a chemical, enzymatic or physical action indicative of the presence or absence of a particular interesting polynucleotide sequence. An example of a "reaction" is an amplification reaction such as a polymerase chain reaction (PCR). Another example of a "reaction" is a sequencing reaction, either by synthesis, ligation, hybridization or by passing DNA through a pore and measuring signals indicative of a particular nucleotide. An "informative response" is a response that indicates the presence of one or more particular interesting polynucleotide sequences, and in a case where only one interesting sequence is present. The term "well" as used herein refers to a response at a predetermined location within a limited structure, e.g., a well-shaped vial, cell, or chamber in a PCR array or e.g. the individual reaction volumes in which sequencing reactions take place (including so-called cartridge flow cells from Illumina).

The term "clinically relevant nucleic acid sequence" or "target chromosome or chromosomal segment" as used herein may refer to a polynucleotide sequence corresponding to a segment of a larger genomic sequence whose potential imbalance has been tested or to the larger genomic sequence itself. An example is the sequence of chromosome 21. Other examples include chromosome 18, 13 X and Y. Still other examples include mutated genetic sequences or genetic polymorphisms or copy number variations (CNVs) that a fetus can inherit from one or both of the parents. Still other examples include sequences that have been mutated, deleted, or amplified in a malignant tumor, e.g., sequences in which loss of heterozygosity or gene duplication occurs. In some embodiments, multiple clinically relevant nucleic acid sequences, or equivalent multiple makers of the clinically relevant nucleic acid sequences, may be used to provide data for detecting the imbalance. For example, data from five non-consecutive sequences on chromosome 21 can be used in an additional way to determine possible imbalance in chromosome 21, effectively reducing the need for sample volume to 1/5.

The term "over-represented nucleic acid sequence" as used herein refers to the nucleic acid sequence of two interesting sequences (e.g., a clinically relevant sequence and a background sequence) that is present in more abundance than the other sequence in a biological sample.

The term "based on" as used herein means "based at least in part" and refers to one value (or result) that is used in determining another value, such as occurs in connection with an input to a method and the output of that method. The term "derivation" as used herein refers to the relationship of an input of a method and the output of that method, as occurs when the derivation is the calculation of a formula.

The term "parameter" herein refers to a numerical value that characterizes a quantitative data series and / or a numerical relationship between quantitative data series.

The term "score" as used herein refers to a numerical value associated with or based on a specific characteristic, e.g., the number of readings or reading counts for a particular sequence present in a sample. The term "first score" is used herein to refer to a numerical value associated with the target chromosome or chromosomal segment. Another example of a score is, for example, a Z score that quantifies how much the number of readings of a particular sequence differs from the number of readings obtained from the same sequence in a series of reference samples. It is well known to a person skilled in the art how such a Z-score can be calculated.

The term "threshold" or "threshold" as used herein means a numerical value whose value is used to distinguish between two or more statuses (e.g., sick and non-disease) of classification for a biological sample. For example, if a parameter is greater than the threshold value, a first classification of the quantitative data is made (e.g., disease status); or if the parameter is lower than the threshold value, a different classification of the quantitative data is made (e.g., non-disease status).

The term "imbalance" as used herein means any significant deviation as defined by at least one threshold in an amount of the clinically relevant nucleic acid sequence of a reference amount. For example, the reference amount could be a ratio of 3/5, and an imbalance could then occur if the measured ratio is 1: 1.

The term "random sequencing" as used herein refers to sequencing, wherein the sequenced nucleic acid fragments are not specifically identified or intended before the sequencing procedure. Sequence-specific primers to target specific gene loci are not required. The groups of sequenced nucleic acids vary from sample to sample and even from analysis to analysis for the same sample. The identities of the sequenced nucleic acids are only disclosed from the generated sequencing output. In some embodiments of the present invention, random sequencing can be preceded by procedures for enriching a biological sample with certain populations of nucleic acid molecules that share certain common features. In one embodiment, each of the DNA fragments in the biological sample have an equal chance of being sequenced.

The term "fraction of the human genome" or "portion of the human genome" as used herein refers to less than 100% of the nucleotide sequences in the human genome comprising about 3 billion base pairs of nucleotides. In the context of sequencing, it refers to less than 1-fold coverage of the nucleotide sequences in the human genome. The term can be expressed as a percentage or absolute number of nucleotides / base pairs. As a user example, the term can be used to refer to the actual amount of sequencing that has been performed. Embodiments can determine the required minimum value for the sequenced fraction of the human genome to obtain an accurate diagnosis. As another use example, the term may refer to the amount of sequenced data used to derive a parameter or amount for the classification of diseases.

The term "summary statistics" as used herein is used as a statistical term, and refers to an indication of the extent of a distribution of values or scores, or an indication of the score / value present at the center of the distribution. This can be, for example, an average or median or standard deviation (StDev) or median absolute deviation (mad) or average absolute deviation of a set of scores.

The term "fetal fraction" as used herein refers to the fraction of fetal nucleic acids present in a sample comprising fetal and maternal nucleic acid.

The term "copy number variation" or "CNV" (copy number variation) herein refers to variation in the number of copies of a nucleic acid sequence that is a few bp or larger present in a test sample compared to the copy number of the nucleic acid sequence present in a qualified sample . A "copy number variant" refers to the single bp large or larger nucleic acid sequence in which differences in copy numbers are found by comparing an interesting sequence in the test sample with those present in a qualified sample. Copy number variants / variations include deletions, including microdeletions, insertions, including micro-insertions, duplications, multiplications. CNVs include chromosomal aneuploidies and partial aneuploidies.

The term "aneuploidy" herein refers to an imbalance of genetic material caused by the loss or enhancement of a complete chromosome, or part of a chromosome. Aneuploidy refers to both chromosomal and subchromosomal imbalances, such as, but not limited to, deletions, microdeletions, insertions, micro-insertions, copy number variations, duplications. Copy number variations can vary in size in the range from a few bp to several Mb, or in special cases from 1 kb to several Mb. Large subchromosomal aberrations that extend over dozens of MBs and / or correspond to a significant portion of a chromosome arm can also be called segmental aneuploidies.

The term "chromosomal aneuploidy" herein refers to an imbalance of genetic material caused by the loss or enhancement of a complete chromosome, and includes germline anuploidy and mosaic anuploidy.

The term "partial aneuploidy" herein refers to an imbalance of genetic material caused by a loss or enhancement of a portion of a chromosome, e.g., partial monosomy and partial trisomy, and includes imbalances resulting from translocations, deletions, and insertions.

The term "polymorphism, polymorphic target nucleic acid", "polymorphic sequence", "polymorphic target nucleic acid sequence" and "polymorphic nucleic acid" are used interchangeably to refer to a nucleic acid sequence containing one or more polymorphic sites.

The term "polymorphic site" herein refers to a single nucleotide polymorphism (SNP, a single nucleotide polymorphism), a small-scale multi-base deletion or insertion, a Multi-Nucleotide Polymorphism (MNP) or a Short Tandem Repeat (STR, short tandem repeats) or a CNV (copy number variation).

The term "multiple" is used herein with reference to a number of nucleic acid molecules or sequence tags or readings sufficient to identify significant differences in copy number variations (e.g., chromosome doses) in test sample and qualified samples using the methods of the invention. In some embodiments, at least about 3x106 sequence tags, at least about 5x10E6 sequence tags, at least about 8x10E6 sequence tags, at least about 10x10E6 sequence tags, at least about 15x10E6 sequence tags, at least about 20x10E6 sequence tags, at least about 30x100E6 at least 6x10E0E6 sequence tags sequence tags, or at least about 50x10E6 sequence tags obtained for each test sample. Each sequence tag can be a single sequence reading from 20 to 400 bp, or a pair of 2 sequence readings with a paired end of each 20 to 400 bp.

The terms "polynucleotide", "nucleic acid" and "nucleic acid molecules" are used interchangeably and refer to a covalent-linked sequence of nucleotides (ie ribonucleotides for RNA and deoxyribonucleotides for DNA) in which the 3 'position of the pentose of one nucleotide is bound by a phosphodiester group to the 5 'position of the pentose of the following, include sequences in any form of nucleic acid, but not limited to RNA and DNA molecules. The term "polynucleotide" includes, but is not limited to, single and double stranded polynucleotide.

The term "portion," when used with reference to the amount of sequence information of fetal and parent nucleic acid molecules in a biological sample, herein refers to the amount of sequence information of fetal or parent nucleic acid molecules in a biological sample that together are lower in number than the sequence information of <1 human genome.

The term "test sample" herein refers to a sample comprising a mixture of nucleic acids comprising at least one nucleic acid sequence suspected of having undergone copy number variation or at least one nucleic acid sequence for which it is desirable to determine whether a copy number variation exists. Nucleic acids present in a test sample are called test nucleic acids, or target nucleic acids or target chromosomes or target chromosomal segments.

The term "reference sample" herein refers to a sample comprising a mixture of nucleic acids for which the sequencing data is used together with the sequencing data of the test sample for calculating scores and parameters as described in claim 1. Although it is not necessary, a reference sample is at preferably normal (ie not aneuploid) for the sequence of interest. Thus, a reference sample is preferably a qualified sample that does not carry a trisomy 21 and that can be used to identify the presence of a trisomy 21 in a test sample.

The term "reference set" includes several "reference samples".

The term "enrichment" herein refers to the process of specifically amplifying certain target nucleic acids contained in a portion of a sample from the mother. The amplified product is then often combined with the remainder of the sample from the mother from which the portion was removed.

The term "interesting sequence" herein refers to a nucleic acid sequence that is associated with a difference in sequence representation in healthy versus diseased individuals. An interesting sequence may be a sequence on a chromosome that is misrepresented, i.e., over- or under-represented, in a disease or genetic disorder. An interesting sequence can also be a part of a chromosome, or a chromosome. For example, an interesting sequence may be a chromosome that is overrepresented in an aneuploidy disorder. Interesting sequences include sequences that are over- or under-represented in the total population, or a sub-population of cells from a patient.

The term "group of chromosomes" herein refers to two or more chromosomes. The term "set" refers to a set of chromosomes or chromosomal segments, but can also refer to a set of values or scores derived from a corresponding set of chromosomes or chromosomal segments.

The term "reading" refers to an experimentally obtained DNA sequence that is sufficiently long (e.g., at least about 20 bp) that can be used to identify a larger sequence or region, e.g., that can be aligned and in particular assigned to a chromosome location or genomic region or gene. The terms "reading" and "sequences" can be used interchangeably in the draft version.

The term "reading theorem" refers to the number of readings retrieved from a sample assigned to a reference genome or part of said reference genome (bin).

The term "piece" (bin) of a genome is to be understood as a segment of the genome. A genome can be subdivided into different pieces, with either a fixed or a predetermined size or a variable size. A possible fixed piece size can be, for example, 10 kB, 20 kB, 30 kB, 40 kB, 50 kB, 60 kB, 70 kB, etc., where kB stands for kilobase pairs, a unit corresponding to 1000 base pairs.

The term "window" is to be understood as multiple pieces.

The terms "aligned", "alignment", "assigned" or "alignment", "assignment" refers to one or more sequences identified as a match in terms of the order of their nucleic acid molecules with a known sequence of a reference genome. Such alignment can be done manually or by a computer algorithm, examples of which include the Efficient Local Alignment or Nucleotide Data (ELAND) computer program as part of the Illumina Genomics Analysts pipeline. The correspondence of a sequence reading in the alignment may be a sequence match of 100% or less than 100% (non-perfect match).

The term "reference genome" as used herein refers to a digital nucleic acid sequence database, assembled as a representative example of a kind of DNA. Since it is assembled from the sequencing of multiple DNA, a reference genome does not accurately represent the DNA of a single person. It is used to allow the assignment of sequencing readings from a sample to specific chromosomal positions.

The term "clinically relevant sequence" herein refers to a nucleic acid sequence that is well known and suspected of being associated or associated with a genetic or disease condition. Determining the absence or presence of a clinically relevant sequence may be useful in determining a diagnosis or confirming a diagnosis of a medical condition, or making a prognosis for the development of a disease.

The term "derived" when used in the context of a nucleic acid or a mixture of nucleic acids, refers herein to the means by which it or the nucleic acids are obtained from the source from which they originate. In one embodiment, a mixture of nucleic acids derived from two different genomes means, for example, that the nucleic acids, e.g., cell-free DNA, were naturally released by cells by naturally occurring processes such as necrosis or apoptosis, or by lysis of the cells due to improper storage. or transportation conditions.

The term "mother sample" herein refers to a biological sample obtained from a pregnant patient, e.g., a woman.

The term "biological fluid" herein refers to a fluid taken from a biological source and includes, for example, blood, serum, plasma, sputum, wash fluid, cerebrospinal fluid, urine, sperm, sweat, tears, saliva, blastocoel fluid, and the like. It also refers to the medium in which biological samples can be grown, such as in vitro culture medium in which cells, tissue or embryo can be grown. As used herein, the terms "blood", "plasma" and "serum" include fractions or processed portions thereof. Similarly, when a sample is taken from a biopsy, smear, smear, etc., the "sample" explicitly includes a processed fraction or portion derived from the biopsy, smear, smear, etc.

The terms "mother nucleic acids" and "fetus nucleic acids" herein refer to the nucleic acids of a pregnant female patient and the nucleic acids of the fetus carried by the pregnant woman, respectively. As explained above, "fetal nucleic acids" and "placenta nucleic acids" are often used to refer to the same type of nucleic acids, although there may be biological differences between the two types of nucleic acids.

The term "corresponding" herein refers to a nucleic acid sequence, e.g., a gene or a chromosome, that is present in the genome of different patients, and that does not necessarily have the same sequence in all genomes, but serves to provide identity rather than the genetic information of an interesting sequence, e.g. a gene or chromosome.

The term "substantially cell-free" herein refers to preparations of the desired sample from which components that are normally associated with them have been removed. For example, a plasma sample can be made substantially cell-free by removing blood cells, e.g., white blood cells, which are normally associated with it. In some embodiments, essentially free samples are processed to remove cells that would otherwise contribute to the genetic material to be tested for aneuploidy.

As used herein, the term "chromosome" refers to the hereditary gene carrier of a living cell derived from chromatin and comprising DNA and protein components (in particular histones). The conventionally internationally recognized individual human genome chromosome numbering system is used herein. The term "chromosomal segments" is to be understood as a part of a chromosome. Said segments may refer to a part, window, or specific region within a chromosome, e.g., known to include, for example, deletions or insertions or copy number variations.

As used herein, the term "polynucleotide length" refers to the absolute number of nucleic acid molecules (nucleotides) in a sequence or in a region of a reference genome. The term "chromosome length" refers to the known length of the chromosome given in base pairs.

The term "patient" herein refers to a human patient as well as a non-human patient such as a mammal, an invertebrate animal, a fungus, a yeast, a bacterium and a virus. Although the examples herein refer to human genomes and the description is primarily directed to humans, the concept of the present invention applies to genomes of any plant or animal, and can be used in the field of veterinary medicine, animal sciences, research laboratories and the like.

The term "condition" here refers to "medical condition" as a broad term that covers all diseases and conditions, but which may include injuries and normal healthy situations, such as pregnancy, which may affect a person's health, benefit from have medical assistance or implications for medical treatments.

The term "attribute" is to be understood as a property or value of an object or element. This can be, for example, a certain correction factor that is used to correct the reading theorem for a particular polymorphism. An attribute can be defined experimentally using a series of samples.

DETAILED DESCRIPTION

The invention relates to a method for determining the presence or absence of a fetal chromosomal aneuploidy in a pregnant woman. This determination can be performed by calculating a parameter associated with chromosomal data obtained from a biological sample. A computer-readable medium is also provided which is coded with multiple instructions for controlling a computer system for performing the methods.

In a second aspect, the present invention describes a methodology for determining the fetal fraction in a sample. In particular, the method allows the determination of the fraction of cell-free DNA (cfDNA) that is contributed by a fetus to the mixture of fetal and maternal cfDNA in a sample from the mother, e.g., a plasma sample. In a preferred embodiment, the present invention permits both the determination of the presence or absence of a fetal chromosomal aneuploidy in a pregnant woman and the determination of the fetal fraction independently of the gender of the fetus.

In one aspect, reading counts are determined based on the sequencing of nucleic acid molecules in a sample from the mother, such as urine, plasma, serum, blastocoel fluid, and other suitable biological samples. Nucleic acid molecules from the biological sample are randomly sequenced so that a fraction of the genome is sequenced. One or more threshold values are selected to determine whether there is a change compared to a reference amount (i.e., imbalance), for example, with respect to the ratio of amounts of two chromosomal regions (or arrays of regions).

The detected change in the reference amount can be any deviation (up or down) with respect to the clinically relevant nucleic acid sequence or the target chromosome or chromosomal segment relative to the other non-clinically relevant sequences. Thus, the reference status can be a ratio or other amount (e.g., other than a 1-1 match), and the measured status indicating a change can be a ratio or other amount that differs from the reference amount as determined by one or more thresholds.

The clinically relevant chromosomal region (also referred to as a clinically relevant nucleic acid sequence or target chromosome or chromosomal segment) and the background nucleic acid sequence may be from a first type of cells or from one or more second types of cells. For example, fetal nucleic acid sequences from cells of the fetus / placenta may be present in a biological sample, such as maternal plasma, which contains a background of maternal nucleic acid sequences from maternal cells. Note that the percentage of fetal sequences in a sample can be determined by any fetus derived loci and is not limited to measuring the clinically relevant nucleic acid sequences. 1. General method for evaluating an aneuploidy The present invention describes a methodology for detecting the presence or absence of a fetal chromosomal aneuploidy and / or the determination of the fetal fraction present in a biological sample.

In a first aspect, the method for detecting the presence or absence of a fetal chromosomal aneuploidy is based on the determination of a parameter from the nucleic acid content of a biological sample. The biological sample can be plasma, urine, serum, blastocoel fluid or any other suitable sample. The sample contains nucleic acid molecules from the fetus and the pregnant woman. The nucleic acid molecules can be, for example, fragments of chromosomes.

At least a portion of a plurality of the nucleic acid molecules included in the biological sample is randomly sequenced to obtain a plurality of sequences. The sequenced portion represents a fraction of the human genome and can be isolated from the sample by conventional means (e.g., cell-free DNA extraction and preparation of an NGS library). In one embodiment, the nucleic acid molecules are fragments of respective chromosomes. One end (e.g., 50 base pairs (bp)), both ends, or the entire fragment can be sequenced. A subset of the nucleic acid molecules in the sample can be sequenced, and this subset is randomly selected, as will be described in more detail below.

In one embodiment, the random sequencing is done using solid parallel sequencing. Solid parallel sequencing, such as this achieved on the HiSeq2000, HiSeq2500, HiSeq3000, HiSeq4000, HiSeq X, MiSeq, MiSeqDx, NextSeq500, NextSeq550 flowcell, the 454 platform (Roche), Illumina Genome Analyzer (or SoleiDysystem) (App SoleiDysystem) ) or PGM or Proton platform (IonTorrent) or GeneRead (Qiagen) or the Helicos True Single Molecule DNA sequencing technology, the single molecule, real-time (SMRT ™) technology from Pacific Biosciences, and nanopore sequencing as in MinlON, PromethION, GridlON (Oxford Nanopore technologies), allow the sequencing of many nucleic acid molecules isolated in parallel from a specimen at higher orders of multiplexing. Each of these platforms sequencially clonally expanded or even non-amplified single molecules of nucleic acid fragments. Clonal expansion can be achieved by bridge amplification, emulsion PCR or Wildfire technology.

Since a large number of sequencing readings, in the order of hundreds of thousands to millions or even possibly one hundred million or billions, are generated from each sample in each run, the resulting sequenced readings form a representative profile of the mix of nucleic acid species in the original specimen. For example, the (De?) Halotype, transcriptome and methylation profiles of the sequenced readings resemble those of the original specimen. Because of the large sampling of sequences from each specimen, the number of identical sequences, such as those generated from the sequencing of a nucleic acid group at different multiples of coverage or high redundancy, is also a good quantitative representation of the count of a particular nucleic acid species or locus in the original sample.

Based on the sequencing (e.g., data from the sequencing), a first score of a target chromosome or chromosomal segment (e.g., the clinically relevant chromosome) is determined. The first score is determined based on sequences identified as coming from (i.e., aligning with) the target chromosome or chromosomal segment. A bio-information procedure can then be used, for example, to locate any of these DNA sequences for the human genome or a reference genome. It is possible that some of such sequences will be removed from later analysis because they are present in the repeat regions of the human genome, or in regions that are subjected to inter-individual variations, e.g., copy number variations. A score of the target chromosome or chromosomal segment and of one or more other chromosomes can thus be determined.

Based on the sequencing, a set of scores from one or more chromosomes or chromosomal segments is determined from sequences identified as coming from (i.e., aligning with) a series of one or more chromosomes. In one embodiment, said set contains all other chromosomes in addition to the first (i.e., the first tested). In another embodiment, said set contains a single different chromosome. In a most preferred embodiment, said set contains chromosomes or chromosomal segments and comprises the target chromosome or chromosomal segment.

There are a number of ways to determine a score. Said score is preferably based on the reading counts obtained from sequencing. Said reading counts may include counting the number of sequenced readings, the number of sequenced nucleotides (base pairs) or the accumulated lengths of sequenced nucleotides (base pairs) originating from a particular chromosome (s) or chromosomal segments such as pieces or windows or clinically relevant chromosome parts.

Rules can be imposed on the results of the sequencing to determine what is counted. In one aspect, a reading theorem can be obtained based on a part of the sequenced output. For example, sequencing output corresponding to nucleic acid fragments with a specified size range could be selected.

In one embodiment, said score is the raw reading count for a particular chromosome or chromosomal segment.

In a preferred embodiment, said reading counts are subjected to mathematical functions or operations to derive said score from said reading count. Such operations include, but are not limited to, statistical operations, regression models, standard calculations (addition, subtraction, multiplication, and division), said standard calculations preferably being based on one or more reading counts obtained.

In a preferred embodiment, said first score is a normalized value derived from the reading counts or mathematically modified reading counts. In another preferred embodiment, said score is a Z score or standard score with respect to the reading counts of a particular chromosome, chromosomal segment or the mathematically modified counts thereof, the Z score quantifying how much the number of readings of a given sequence differs of the number of readings obtained from the same sequence in a series of reference samples. It is well known to a person skilled in the art how such a Z-score can be calculated.

In a preferred embodiment, a parameter is determined based on a first score (corresponding to the target chromosome or chromosomal segment) and a set of scores. The parameter preferably represents a relative score between the first score and a summary statistical of the set of scores. The parameter may, for example, represent a simple ratio of the first score to a summary statistic of the set of scores. In one aspect, each score could be an argument of a function or individual functions, whereby a ratio can then be taken of these individual functions.

In a preferred embodiment, the parameter can be obtained by a ratio between: - a first function where the first score and the set of scores are the arguments; - a second function where the set of scores is the argument.

In a preferred embodiment, said first function is defined as a difference, preferably the difference between the first score and a summary statistic of the set of scores, said summary statistic preferably being selected from the mean, the median, the standard deviation or median absolute deviation (mad) or average absolute deviation.

In another preferred embodiment, said second function is defined as a variability summary statistic of the set of scores, said summary statistic preferably being selected from the mean, median, standard deviation or median absolute deviation (mad) or average absolute deviation .

A suitable embodiment of the present invention usually comprises the following steps (after obtaining DNA sequences from a random, low-coverage sequencing process on a biological sample). - aligning sequences with a reference genome; - obtaining the reading counts per chromosome or chromosomal segment; - normalizing the number of readings or a derivative thereof to a normalized number of readings; - obtaining a first score derived from said normalized number of readings and a set of scores derived from said normalized reading counts for a target chromosome or chromosomal segment, and wherein said set of scores is a set of scores derived from the normalized number of readings obtained from a series of chromosomes or chromosome segments comprising the chromosomal target segment or chromosome; - calculating a parameter of said scores, wherein said parameter represents a ratio between said first score and a summary statistic of said set of scores, the first function of said ratio being defined as a difference between the first score and a summary statistic of said set of scores; and wherein the second function of said ratio is defined as a summary statistic of said set of scores.

Said sequences are preferably obtained by low-coverage sequencing.

Said normalization preferably takes place on the basis of a series of reference samples, said reference samples preferably being, but not necessarily, euploid or substantially euploid for the chromosome or chromosomal segment corresponding to the target chromosome or chromosomal segment (ie, the major part of the chromosome or chromosomal segment in the reference samples corresponding to the target chromosome or chromosomal segment in the test sample are euploid). Such a reference series has different sample sizes. A possible sample size can be, for example, 100 samples, such as 50 male and 50 female samples. It will be clear to a person skilled in the art that the reference series can be freely chosen by the user.

Said number of readings is preferably recalibrated to correct for GC content and / or total number of readings obtained from said sample.

By taking into account a series of scores derived from readings from chromosomes or chromosomal segments that contain the target chromosome or chromosomal segment, a more sensitive and reliable parameter can be obtained in comparison with prior art methods. Unlike the methods known in the art, no assumption has to be made about the ploidy status of any of the chromosomes in the test sample. Namely, by defining a parameter according to the present invention, the parameter for the chromosome or region to be analyzed is clear (i.e., it is greatly increased / decreased) and does not disappear in the noise (i.e., only moderately or not increased / decreased). Sensitivity is also essential for screening, since it is important to have a reliable and reliable result, minimizing the number of false negatives. For screening it can be more important to have high superiority compared to specificity.

The parameter according to the present invention allows robust detection and automatic classification of chromosomes, even in data with noise. By taking into account a set of chromosomes or segments, including the target chromosome or segment, i.e., the majority of information present in the data set, the majority of the available information is used, resulting in a more adequate analysis. If, for example, chromosome 1 (the largest chromosome, 7.9% of the genome) were to be deleted, a large amount of data would be deleted which would not be taken into account, which would cause a disruption in the analysis.

The present invention is particularly useful in situations where a low number of readings or data with noise is obtained. The inventors have found that, in the latter situations, the parameter of the present invention was superior to other methodologies.

In a preferred embodiment, said scores are obtained based on the genomic representation of the target chromosome or chromosomal segment (or a region thereof) and the genomic representation of all autosomes or chromosomes, thereby recording the target chromosome or chromosomal segment.

The parameter is compared with one or more threshold values. The threshold values can be determined in any number of suitable ways. Such ways include Bayesian-type probability method, sequential probability test (SPRT), discovery of false results, confidence interval, receiver-operated characteristic (ROC, receiver operating characteristic). In a more preferred embodiment, said threshold value is based on statistical considerations or is determined empirically by testing biological samples. The threshold value can be validated by means of test data or a validation series and, if necessary, can be changed whenever more data is available.

It is possible that in some variants of the procedure, the threshold value would be adjusted in accordance with information on the cell-free fetal DNA fraction in the mother's plasma sample (also called fetal fraction or abbreviated ff or f). In another embodiment, said fetal fraction can serve as an internal control of the quality of the sample. The value of f can be determined in various ways based on the sequencing data set (depending on the gender of the fetus, or independently of the gender of the fetus), as will be further explained below.

Based on the comparison, a classification is determined as to whether there is a fetal chromosomal aneuploidy for the target chromosome or chromosomal part. In one embodiment, the classification is a definitive yes or no. In another embodiment, the classification may not be classifiable or uncertain. In yet another embodiment, the classification may be a risk score to be interpreted at a later date, for example, by a physician.

In a further preferred embodiment, secondary parameters of the reading counts are calculated, which serve as an additional internal check for the usability of the parameter, the extent of the aneuploidy (if identified) and / or an indication for the reliability of the parameter, the biological sample or the sequences obtained therefrom and therefore the final assessment. The value for said secondary parameters can be, for example, a measure or requirement of the presence of said aneuploidy and / or a measurement of the quality of the sample.

In one embodiment, said secondary parameter is calculated as the median of the Z distribution of the reading counts or a derivative thereof, for a target chromosome or chromosomal target segment measured per piece or a set of pieces (i.e., windows). The last secondary parameters allow assessment if the majority (more than 50%) of the windows in a chromosome have increased or decreased. The latter allows the detection of chromosomal and large subchromosomal aneuploids. If less than 50% of the windows are affected, the secondary parameters will not be affected (eg smaller CNVs).

In another embodiment, said secondary parameters can be calculated as the median of the absolute value of the Z-scores for the reading counts or a derivative thereof, of the remaining chromosomes (that is a collection of chromosomes or segments representing the target chromosome or chromosomal segment exclude).

The latter secondary parameters allow the detection of, inter alia, the presence of technical or biological instabilities and the distinguishing thereof from CNVs of the mother. If fewer than the windows of the other of all chromosomes are affected, these secondary parameters will not be affected. If more than 50% of the windows are affected, this can be deduced from the aforementioned secondary parameters.

In another embodiment, the present invention also provides a quality score (QS). QS makes it possible to assess the general variation within the genome. A low QS is an indication of good sample processing and a low level of technical and biological noise. An increase in the QS can have two possible reasons. Either an error occurred during the processing of the sample. In general, the user will be asked to take and test a new biological sample. This is typical of moderately increased QS scores. A greatly increased QS could be an indication of a strong aneuploid sample and the user will be encouraged to do a confirmatory test. Said QS is preferably determined by calculating the standard deviations of all Z scores for chromosomes or chromosomal segments and optionally by removing the outliers thereof (i.e. the highest and lowest Z scores in this set).

As an alternative or additional embodiment, the determination of the fetal fraction (see below) also serves as an internal quality check of the sample and the sequences obtained therefrom. The quality of a sample can be disturbed after collection, for example due to unsuitable circumstances during collection, transport or storage. The latter can have an effect on the cell-free DNA in the sample, for example due to rupture of (maternal) white blood cells. Consequently, the main group of free-floating DNA will be more enriched for maternal DNA, thereby reducing the percentage of the fetal fraction compared to the total cell-free DNA content in the sample. Said fetal fraction will preferably be determined by at least one of the methods described below.

In an embodiment of the present invention, the parameter will be sufficient to distinguish between the presence and / or absence of an aneuploidy. In a more preferred embodiment of the present invention, both the parameter and the secondary parameters will be used to make a decision regarding the presence or absence of an aneuploidy. The said secondary parameters will also preferably be compared with predefined threshold values.

The methodology of the present invention is preferably particularly suitable for analyzing aneuploidies associated with segments or deletions given in Table 1 that contains a non-exhaustive list of chromosome aberrations that may be identified by the methods and kits described herein . In another or further embodiment, said chromosomal target segment is selected from a piece or window derived from chromosome X, Y, 6, 7, 8, 13, 14, 15, 16, 18, 21 and / or 22.

In a further or different embodiment, said chromosome is selected from chromosome X, Y, 6, 7, 8, 13, 14, 15, 16, 18, 21 and / or 22.

Table 1

II Seauencina. uitliinina and correction

As mentioned above, only a fraction of the genome is sequenced. In one aspect, even when a group of nucleic acids in a specimen is sequenced at <100% genomic coverage rather than with multiple multiples of coverage, and from the ratio of sequenced nucleic acid molecules, most of each nucleic acid species is not sequenced or only sequenced once.

This contrasts with situations where targeted enrichment is performed on a subset of the genome prior to the sequencing reaction, followed by high-coverage sequencing of that subset.

In one embodiment, said sequences are obtained by next generation sequencing. In another preferred embodiment, said sequencing method is a low-coverage, random sequencing method.

In one embodiment, massive parallel short reading sequencing is used. Short sequence tags or readings are generated, e.g. from a certain length between 20 bp and 400 bp. Paired end sequencing could also be performed.

In one embodiment, a pre-processing step is available for pre-processing the obtained readings. Such prior processing option allows filtering of low quality readings, thereby preventing them from being assigned. Assignment of low-quality readings may require lengthy computer processing capacity, may be incorrect, and has the risk of increasing the technical noise in the data, resulting in a less accurate parameter. Such prior processing is particularly valuable when next-generation sequencing data is used that has a generally lower quality or any other condition associated with a generally lower quality of the readings.

The generated readings can be later aligned with one or more human reference genome sequences. The number of aligned readings are preferably counted and / or sorted according to their chromosomal location.

An additional cleaning protocol can be performed, whereby deduplication is performed, e.g. with Picard instruments, where only uniquely assigned readings are retained. Readings with mismatches and gaps can be removed. Readings that divide the blacklisted areas can be excluded. Such blacklisted areas can be taken from a predefined list of, for example, common CNVs, collapsed repeats, DAC blacklisted areas as identified in the ENCODE project (ie a set of areas in the human genome that have abnormal, unstructured, high has signal / reading counts in NGS experiments independent of cell line and type of experiment) and the undefined parts of the reference genome. In one embodiment, areas on the blacklist are given to the user. In another embodiment, the user can use or define his or her own set of areas on the blacklist.

In another embodiment, chromosomes are subdivided into regions with a predefined length, generally referred to as pieces. In one embodiment, the piece size is a predefined size given to the user. In another embodiment, said piece size may be defined by a user, it may be uniform for all chromosomes, it may be a specific piece size per chromosome, or it may vary according to the obtained sequence data. Changing the piece size can have an effect on the final parameter to be defined, either by improving the sensitivity (usually obtained by reducing the piece size, often at the expense of specificity) or by improving specificity (in generally by increasing the piece size, often at the expense of sensitivity). A possible piece size that provides acceptable specificity and sensitivity is 50 kb.

In a further step, the aligned and filtered readings are counted within one piece to obtain reading counts.

The obtained reading counts can be corrected for the GC count for the piece. GC bias is known to exacerbate genome assembly. Various GC corrections are well known in the art (e.g., Benjamini et al., Nucleic Acid Research 2012). In a preferred embodiment, said GC correction will be a LOESS regression. In one embodiment, a user of the methodology of the present invention may be provided with the choice of various possible GC corrections.

In a later step, the genomic representation (GR, genomic reproduction) of reading counts is calculated. Such representation is preferably defined as a ratio between the GC-corrected reading counts for a specific piece and the sum of all GC-corrected reading counts.

In one embodiment, said GR is defined as follows:

with k over all chromosomal pieces.

The factor 107 (or 10E7) in the above formula is arbitrarily defined, and can be any constant value.

In a final step, the obtained GC are joined piece by piece over an area, said area being a sub-area (window) of a chromosome or the entire chromosome. Said window may have a predefined or variable size, which may be optionally selected by the user. A possible window could have a size of 5 MB or 100 adjacent pieces with a size of 50 kb.

The GR pooled for a chromosome can be defined by

In another embodiment, the genomic representation of a series of reference samples must be calculated. Said set of reference samples (or also referred to as reference series) can be predefined or selected by a user (e.g. selected from his / her own reference samples). By allowing the user to use his own reference set, a user will be able to better capture the recurring technical variation of his / her environment and its variables (eg different wet lab reagents or protocol, different NGS instrument or platform, etc.) . In a preferred embodiment, said reference set comprises genomic information of 'healthy' samples that are expected or known to contain (relevant) aneuploidies. The genomic representation (GR) of the reference series can be defined either at the level of the genome and / or at a sub-area (chromosome, chromosomal segment, window or bin).

Other single molecule sequencing strategies such as the Roche 454 platform, the Applied Biosystems SOLiD platform, the Hélicos True Single Molecule DNA sequencing technology, the single molecule, real-time (SMRT ™) technology from Pacific Biosciences, and nanopore sequencing technologies such as MinlON, GridlON or PromethION from Oxford Nanopore Technologies could also be used in this application. III Determination of scores, parameter and secondary parameters

Based on the alignments and the obtained reading counts or a derivative thereof, optionally corrected for GC content and / or total number of readings obtained from said samples, scores are calculated that ultimately lead to a parameter that allows the presence of an aneuploidy in a determine the sample. The scores mentioned are normalized values derived from the counts of the readings or mathematically modified counts of the readings, with normalization taking place for the reference set. Consequently, each score is obtained by comparison with the reference series. The term first score is used to refer to the score associated with the count of the readings for a target chromosome or a chromosomal segment. A set of scores is a set of scores derived from a set of normalized number of readings that may include the normalized number of readings from said chromosomal target segment or target chromosome.

Said first score preferably represents a Z score or standard score for a target chromosome or chromosomal segment. Said set is preferably derived from a set of Z scores obtained from a corresponding set of chromosomes or chromosomal segments comprising said chromosomal target segment or target chromosome.

In a most preferred embodiment, the first score and the set of scores are calculated based on the genomic representation of either a target chromosome or a chromosomal target segment, or all autosomes (or regions thereof) including the target chromosome or chromosome segment.

Such scores can be calculated as follows:

With i a window or a chromosome or a chromosome segment.

A summary statistic of the said set of scores can be calculated, for example, as the average or median value of the individual scores. Another summary statistic of the said set of scores can be calculated as the standard deviation or median absolute deviation or average absolute deviation of the individual scores.

The said parameter p will be calculated as a function of the first score and a derivative (e.g. summary statistics) of the set of scores. In a preferred embodiment, said parameter will be a ratio between the first score corrected by the set of scores (or a derivative thereof) and a derivative of the said set of scores.

In another embodiment, said parameter will be a ratio between the first score corrected by a summary statistic of a first set of scores and a summary statistic of another, second set of scores, both sets of scores comprising the first score.

In a specific preferred embodiment, said parameter p is a ratio between the first score corrected by a summary statistic of said set of scores, and a summary statistic of said set of scores. The summary statistic is preferably selected from the mean, the median, the standard deviation, the median absolute deviation or the average absolute deviation. In one embodiment, the said two used summary statistics in the function are the same. In another more preferred embodiment, said summary statistics differ from the set of scores in the numerator and denominator.

A suitable embodiment of the present invention usually comprises the following steps (after obtaining DNA sequences from a random sequencing process on a biological sample). - aligning said obtained sequences with a reference genome; - counting the number of readings on a series of chromosomal segments and / or chromosomes, whereby counts of readings are obtained; - normalizing said counts of readings or a derivative thereof to a normalized number of readings; - obtaining a first score and a set of scores derived from said normalized reading counts for a target chromosome or chromosomal segment, and wherein said set of scores is a set of scores derived from a corresponding set of chromosomes or chromosome segments that comprise the chromosomal target segment or chromosome; - calculating a parameter p on the basis of said first score and said set of scores, wherein said parameter represents a ratio between * said first score, corrected by a summary statistic of said set of scores, and * a summary statistic of the aforementioned set of scores.

A possible parameter p can be calculated as follows:

wherein Zi represents the first score and Z j represents the set of scores and where i represents the target chromosome or chromosomal section, and where j represents a set of chromosomes or chromosomal segments i, a, b, ... representing said chromosomal segment or chromosome i contain.

In another embodiment, the said parameter p is calculated as

Where Z represents the first score and Z j represents the set of scores and where i represents the target chromosome or chromosomal section, and where j represents a set of chromosomes or chromosomal segments i, a, b, ... representing said chromosomal segment or chromosome i.

In yet another most preferred embodiment, said parameter p is calculated as

wherein Z i represents the first score and Z j represents the set of second scores and where i represents the target chromosome or chromosomal section, and wherein j represents a set of chromosomes or chromosomal segments i, a, b, ... representing said chromosomal segment or chromosome i propose.

In addition to the parameter p that allows identification of the presence of an aneuploidy, secondary parameters can be calculated that can serve as a quality check or provide additional information regarding one or more aneuploidies present in the samples.

A first secondary parameter that can be calculated allows defining whether chromosomal and large subchromosomal aneuploidies are present in the sample (compared to, for example, smaller aneuploidies). In a preferred embodiment, such a parameter is defined by a median of Z scores measured per sub-area (e.g., 5 Mb windows) in a target chromosome or chromosomal target section. If more than 50% of these subareas are affected, this will be noticeable in the secondary parameter.

In another embodiment, a secondary parameter can be calculated as the median of the absolute value of the Z scores calculated over the remaining chromosomes (that is, all chromosomes except the target chromosome or chromosomal segment) per sub-area (e.g., 5 Mb windows).

The last secondary parameter allows the detection of the presence of technical or biological instabilities. If less than half the windows of the other or all chromosomes are affected, this secondary parameter will not be affected. If more than 50% of the windows are affected, this can be derived from the aforementioned secondary parameter.

In another embodiment, the present invention also provides a quality score (QS). QS makes it possible to assess the general variation within the genome. A low QS is an indication of good sample processing and a low level of technical and biological noise. An increase in the QS can have two possible reasons. Either an error occurred during the processing of the sample. In general, the user will be asked to take a new biological sample and to sequence it. This is typical of moderately increased QS scores. A greatly increased QS could be an indication of a strong aneuploid sample and the user will be encouraged to do a confirmatory test. Said QS is preferably determined by calculating the standard deviations of all Z-scores for the autosomes or chromosomes and by removing the highest and lowest scoring chromosome.

For example, samples with a QS higher than 2 are considered to be of poor quality, and a QS between 1.5 and 2 is of an intermediate quality. IV. Comparison of threshold value

The parameter p as calculated in the above embodiments will then be compared to a threshold value to determine if there is a change compared to a reference amount (ie imbalance), for example with regard to the ratio of amounts of two chromosomal regions (or series of regions) ). In one embodiment, the user will be able to define his / her own threshold value, either empirically based on experience or previous experiments, or for example based on standard statistical considerations. If a user wants to increase the sensitivity of the test, the user can lower the thresholds (i.e., bring them closer to 0). If a user would like to increase the specificity of the test, the user can raise the thresholds (i.e., move them further from 0). A user will often have to find a balance between sensitivity and specificity, and this balance is often lab and application specific, so it is easy if a user can change the threshold values themselves.

Aneuploidy can be found present or absent based on the comparison with the threshold value.

In an embodiment of the present invention, comparison of parameter p with a threshold value is sufficient to determine the presence or absence of an aneuploidy. In another embodiment, the said aneuploidy is determined on the basis of a comparison of parameter p with a threshold value and a comparison of at least one of the secondary parameters, quality score and / or first score with a threshold value, with a corresponding threshold value for each score. defined or set.

In a preferred embodiment, the said presence / absence of an aneuploidy is defined by comparing a parameter p with a predefined threshold value, as well as comparing all secondary parameters and first scores as described above with their corresponding threshold values.

The final decision tree can thus depend on parameter p alone, or combined with one of the secondary parameters and / or quality score or first score as described above.

In a preferred embodiment, said methodology according to the present invention comprises the following steps: - multiplex sequencing of 50 bp single-end readings (performed by end user) - uploading of sequence readings - assignment of readings to a reference genome - count number of readings per piece (a piece has a size of 50 kb) - calculate GC content per piece and correct for GC content - calculate genomic representation (GR) score per item. For item i this is equal to

wherein GC 1 represents the GC-corrected number of readings for piece i, and GC k represents the GC-corrected number of readings per piece k. combining the GR values per window (a window consists of 100 consecutive windows) calculating a Z score per window or per chromosome, the Z score being based on the GR score per chromosome, compared to the GR score scores in a series of reference samples.

with i a chromosome or a window, μ Ref, j the mean or median GR score for the corresponding pieces in the set of reference samples and σ Ref, i the standard deviation of the GR scores for the corresponding pieces in the set of reference samples - calculation of a ZofZ score, where the ZofZ score is based on the Z score, corrected by the median (or average) of the Z scores of a set of chromosomes or chromosome segments comprising target chromosome i and divided by a factor that determines variability measures the Z-scores of a set of chromosomes comprising the target chromosome i (standard deviation of a more robust version thereof, such as, for example, the median absolute deviation or mad). - comparison of the Z score with a threshold value, and the ZofZ score with a threshold value, for predicting the presence or absence of aneuploidy.

In another preferred embodiment, said prediction of the presence or absence of an aneuploidy takes place via a decision tree based on a parameter p and secondary parameters. V. Determination of fetal fraction

Separately from or in addition to determining the presence of aneuploidy, the present invention also provides one or more methodologies for determining the fetal fraction of fetal nucleic acids in a sample that is a mixture of fetal and maternal nucleic acids. The fetal fraction within the sample will generally be so low that it cannot be easily determined. The fetal fraction in samples generally ranges from 4 to 20% between different samples, with an average of about 10% of the total genomic fraction.

The present invention offers two different methodologies for defining the fetal fraction in a sample, depending on the nature of the pregnancy.

A first methodology is independent of the type of pregnancy or gender of the fetus. Such methodology is based on the presence of polymorphisms in the DNA of the mother (and of the fetus) of the sample. More specifically, it is based on predefined pieces of DNA that may be present in the fetal DNA and absent in the mother's DNA. Polymorphic sites included in the target nucleic acids include, but are not limited to, single-nucleotide polymorphisms (SNPs), tandem SNPs, small-scale multi-base deletions or insertions, said IN-DELs (also called deletion insertion polymorphisms or DIPs), multi -nucleotide polymorphisms (MNPs), copy number variations (CNVs) and short tandem repeats (STRs). In a most preferred embodiment, the polymorphisms are CNVs. In a first step, the sample is sequenced as described above and readings are mapped against a reference genome.

In a later step, the number of sequences that align with each of a predetermined set of polymorphisms is counted. Optionally, this can be achieved by assigning the obtained sequence readings to each of said predefined polymorphisms. Said predefined set of polymorphisms are understood as a set of polymorphisms that have been identified and are believed to be relevant for the determination of the fetal fraction.

The said set of polymorphisms are preferably benign polymorphisms that often occur in the population. In a preferred embodiment, said set of CNVs comprises. Such CNVs can be variable in size, in a preferred embodiment, said CNVs have a length between 10 kb and 1 Nb, more preferably between 10 kb and 100 kb. Alternatively, said CNVs are also 2 bp and 10 Mb long. Herein, kb refers to kilobase pairs (i.e., 1000 base pairs) and Mb refers to megabase pairs (i.e., 1000000 base pairs).

In another preferred embodiment, said set of polymorphisms comprises additional data associated with the polymorphisms within said set. In a preferred embodiment, said data includes one or more attributes of each polymorphism. Said attributes may include, but are not limited to, a correction factor for any polymorphisms, said correction factor providing a link between counts of readings for said polymorphism and the actual fetal fraction. The latter allows a simple correction of the readings obtained according to a polymorphism within a sample, thereby obtaining an estimate of the fetal fraction or the actual fetal fraction.

Said attributes may include a threshold value per polymorphism that allows to identify whether said polymorphism, if present in a sample, qualifies as informative polymorphism.

In a preferred embodiment, the said attributes are determined using a series of values for which the fetal fraction is known. These samples could be, for example, pregnancies of boys (for which the fetal fraction can be determined using readings from the X or Y chromosome, as described below), or else could be samples for which the fetal fraction was determined using orthogonal methods (based on epigenetics or targeted sequencing or digital PCR).

In one embodiment, said set of polymorphisms and attributes are predefined. In another embodiment, said set of polymorphisms and attributes may be defined by the user.

In a later step, the obtained count of readings - or a derivative thereof - is used for each polymorphism to identify whether the particular polymorphism is informative in the sample. Such informative polymorphisms are generally the polymorphisms that have a lower number of readings than a certain threshold value, the threshold value corresponding to a theoretically expected number of readings given the total number of readings for the sample, or a derivative thereof. The lower number of readings observed for such informative polymorphisms is believed to be due to their presence in the fetal genome and not in the mother's genome. The number of readings - or a derivative thereof - should preferably be less than half the theoretically expected number of readings, since this would mean that the polymorphism is not present in 1, 2 or more copies in the mother's genome and therefore only is present in the fetal genome. Based on these informative polymorphisms, an estimate of the fetal fraction can be made by assuming that the obtained count of readings is directly correlated with the fetal fraction. In a preferred embodiment, the fetal fraction is calculated by first correcting the obtained count of readings for each informative polymorphism using its polymorphism attributes, and then taking the median or average of the corrected counts of readings across all informational polymorphisms. A sample must have at least one informative polymorphism to estimate the fetal fraction. The fetal fraction of samples for which no informative polymorphism had been identified can be estimated using alternative methods, provided that the sample was taken from a male fetus (see below).

In one embodiment, the fetal nucleic acid fraction in the fetal and parent nucleic acid mixture is calculated for each of the informative polymorphisms. In a first step, the expected number for the informative polymorphism is determined on the basis of the normalized counts (normalized to e.g. 10,000,000) obtained from said sample. Based on the expected counts (ie the number of readings that would be expected for the polymorphism, given the total number of readings obtained for the sample, and optionally corrected with a polymorphism-specific attribute), an estimate of the fetal fraction of the test sample is then derived .

In one embodiment, this estimate can be determined for each informative polymorphism, using the formula 2 x 100 x observed counts for the informative polymorphism / expected counts for the informative polymorphism.

The expected count is the number of readings that would be expected for the polymorphism, given the total number of readings obtained for the sample, and optionally corrected with a polymorphism-specific attribute. This polymorphism-specific attribute or factor can be derived for any polymorphism in the said set of polymorphisms using a set of samples for which the fetal fraction is known using alternative methods (e.g. for male fetuses using counts of readings on chromosome X or Y).

In one embodiment, the actual fetal fraction can be calculated as follows:

Actual fetal fraction = estimated fetal fraction x factor.

The percentage of fetal fraction is calculated for at least 1, at least 2, at least 3, at least 4, at least 5, at least 6, at least 7, at least 8, at least 9, at least 10, at least 11 , at least 12, at least 13, at least 14, at least 15, at least 16, at least 37, at least 18, at least 19, at least 20, at least 25, at least 30, at least 35, at least at least 40 or more informative polymorphisms. In one embodiment, the fetal fraction will be determined by the average or median fetal fraction as determined by each individual informative polymorphism. In one embodiment, the fetal fraction is the mean or median fetal fraction determined for at least 1, 2 or 3 informational polymorphisms.

The determination of the fetal fraction in a sample from a pregnant woman is generally obtained by obtaining counts of readings from one or more polymorphisms in a sample, and determining whether a polymorphism is informative based on these counts of readings and polymorphism-specific attributes, wherein the count of readings of each informative polymorphism is related to the estimated fetal fraction.

In an embodiment of the present invention, said fetal fraction in a sample is determined as follows: - receiving the sequences of at least a portion of the nucleic acid molecules contained in a biological sample obtained from said pregnant woman; - counting the number of sequences that align with a predefined set of polymorphisms - comparing the obtained number of sequences with the expected number of sequences for each polymorphic site to identify the informative polymorphic site (s) for the sample; - calculating on the basis of the obtained number of sequences for said informative polymorphic site (s) an amount, said amount being an indication of the fetal fraction.

The stated amount is calculated using linear scaling based on informative polymorphism-specific attributes.

In one embodiment, said sequences are obtained by next generation sequencing. In another preferred embodiment, said sequencing method is a random low-coverage, random sequencing method.

The polymorphisms mentioned are preferably copy number variations with a size between 100 bp and 1 Mb, or between 1 kb and 1 Mb, or between 2 bp and 10 Mb.

Said fetal fraction can serve as an internal quality control of the sample and therefore yield another secondary parameter obtained from said sample.

In another embodiment of the present invention, a method for determining the fetal fraction is based on cases where a male pregnancy (i.e., the pregnant woman carries a male fetus) was identified. If a male pregnancy is detected, the fetal fraction can be determined based on readings aligned with the Y chromosome. The X and Y chromosomes usually have regions that are similar between the X and Y chromosomes, called Pseudo-Autosomal Regions (PAR, pseudo-autosomal regions). Since the Y chromosome is small, the influence of these PAR regions is strong, since only a small amount of the readings will be assigned to specific regions within the Y chromosome. The influence of PAR on the X chromosome is less important because of the size of the X chromosome.

In a first step, these regions that are unique to the X or Y chromosome (outside of the PAR regions) are defined. Readings directed against these unique X and / or Y chromosome regions are counted and the fetal fraction is determined based on the unique X and / or Y chromosome regions (or a derivative of these readings, such as normalized readings).

Based on the X chromosomes, the fetal fraction can be determined as twice the difference at the 50 kb piece level between the median number of readings assigned to the autosomes and the median number of readings assigned to chromosome X, divided by the median number of readings assigned to the autosomes. This can be written as the following formula:

Secondly, the fetal fraction can also be estimated based on the Y chromosome since all readings relating to chromosome Y must in theory come from the fetal DNA. The fetal fraction based on chromosome Y is defined as twice the median number of GC-corrected readings assigned to Y over the median number of GC-corrected readings assigned to the autosomes, or in a formula:

The present invention also provides a computer program product comprising a computer-readable medium encoded with a plurality of instructions for controlling a computer system to perform an operation of determining or estimating the fetal fraction in a biological sample obtained from a sample pregnant woman according to the present invention. The processing comprises in particular the following steps: - receiving the sequences of at least a part of the nucleic acid molecules included in a biological sample obtained from said pregnant woman; - counting the number of sequences that align with a predefined set of polymorphisms - comparing the obtained number of sequences with the expected number of sequences to identify the informative polymorphic site (s) for the sample; and - calculating on the basis of the obtained number of sequences for said informative polymorphic site (s) an amount, said amount being an indication of the fetal fraction. VI Gender determination

Sex determination can be done by determining areas that are informative or indicative of male pregnancies. These regions can be defined by looking into a data set comprising sequencing data from male data sets. Areas that are statistically indicative of male pregnancies are retained later.

The readings for one or more informational regions - which usually lie on the Y chromosome - are obtained from the analyzed sample and compared with a predefined threshold value. If the total number of readings across all selected areas in a test sample is higher than a first threshold, it is most likely a male pregnancy. On the other hand, if the total number of readings across all selected areas in a test sample is lower than a second threshold, it is most likely a female pregnancy. If the total number of readings across all selected areas in a test sample is between the first and the second threshold value, the gender cannot be determined (this could be the case of "vanishing twins"). By properly selecting the areas, the first and the second thresholds, the method would be less sensitive to deviations resulting from vanishing twins (especially in the case of boys). VII Toolbox and kit

The methodologies as described above are preferably all implemented by a computer. Therefore, the present invention also relates to a computer program product comprising a computer readable medium encoded with a plurality of instructions for controlling a computer system for performing an operation for performing prenatal diagnosis of fetal aneuploidy and / or screening for fetal aneuploidies and / or the determination of the fetal fraction in a biological sample obtained from a pregnant woman, wherein the biological sample comprises nucleic acid molecules.

With regard to determining the presence or absence of an aneuploidy in a sample, the processing comprises the steps of: - receiving the sequences of at least a portion of the nucleic acid molecules contained in a biological sample obtained from the said pregnant woman; - aligning said obtained sequences with a reference genome; - counting the number of readings on a series of chromosomal segments and / or chromosomes, whereby counts of readings are obtained; - normalizing said counts of readings or a derivative thereof to a normalized number of readings; - obtaining a first score of said normalized readings and a set of scores derived from said normalized readings for a target chromosome or chromosomal segment, and wherein said set of scores is a set of scores derived from the normalized number of readings for a series of chromosomes or chromosome segments comprising the chromosomal target segment or chromosome; - calculating a parameter p based on the said first score and the said set of scores.

In a preferred embodiment, said parameter represents a ratio between: * a first score, corrected by a summary statistic of said set of scores, and * a summary statistic of the set of scores.

With respect to the determination of the fetal fraction in a biological sample, the processing comprises the steps of: - receiving the sequences of at least a portion of the nucleic acid molecules contained in a biological sample obtained from said pregnant woman; - counting the number of sequences that align with a predefined set of polymorphisms - comparing the obtained number of sequences with the expected number of sequences to identify the informative polymorphic site (s) for the sample; and - calculating on the basis of the obtained number of sequences for said informative polymorphic site (s) an amount, said amount being an indication of the fetal fraction.

In another embodiment, said fetal fraction can be determined in the case of a male pregnancy based on the Y chromosome.

Said operations include the steps of - receiving the sequences of at least a portion of the nucleic acid molecules contained in a biological sample obtained from said pregnant woman; - determining the gender of said fetus; wherein if said fetus is male: * aligning said obtained sequences with a reference database; * identifying and counting readings that are specifically in non-PAR regions of the X and / or Y chromosome; * calculating on the basis of said counts of quantity readings, said amount being an indication of the fetal fraction.

Said operations can be performed by a user or professional in an environment away from the location where the sample was taken and / or the wet lab procedure, which is the extraction of the nucleic acids from the biological sample and the sequencing.

The aforementioned operations can be delivered to the user through custom software that must be installed on a computer, and can be stored in the cloud.

After having performed the required or desired operation, the practitioner or user will receive a report or score, the said report or score giving information about the characteristic being analyzed. A report preferably includes a link to a patient or sample ID that has been analyzed. Said report or score provides information about the presence or absence of aneuploidy in a sample, said information being obtained based on a parameter calculated by the above methodology. The report can also provide information about the nature of the aneuploidy (if detected, eg major or minor chromosomal abnormalities) and / or the quality of the sample that has been analyzed.

Said practitioner or user may also receive information about the fetal fraction, said fetal fraction being determined by one of the methodologies of the present invention.

In another embodiment, said practitioner or user may receive information about the sex of the fetus based on the report.

It will be clear to a person skilled in the art that the above information can be presented to a practitioner in one report.

The above operations are preferably part of a digital platform that permits the molecular analysis of a sample through various computer-implemented operations.

In particular, the present invention also includes a visualization tool that allows the user or practitioner to visualize the results obtained as well as the raw data entered into the system. In one embodiment, said visualizations include a window per chromosome, showing the chromosome that has been analyzed, showing readings per area or a score derived therefrom, and the scores and / or parameters calculated. By showing the calculated scores or parameter to the practitioner or user together with the visual representation of the counts of readings, a user can perform an additional check or assessment of the results obtained. By allowing the user to view the data, users will be able to define improved decision rules and thresholds.

In addition, an additional check is added, since the visual data per chromosome allows the user to evaluate for each chromosome whether the automated classification is correct. This adds an additional safety parameter.

In a preferred embodiment, said platform and said visualization instrument are provided with algorithms that take into account the fact that certain areas yield more readings (due to a recurring technical deviation that always makes some areas of the genome over- or under-represented). Correction measurements can be given for this overrepresentation by making a comparison with a reference series (which is ideally processed using the same or a similar protocol) and plotting of, for example, Z-scores or alternative scores that reduce the probability of certain observations under the assumption of propose aneuploidy. Standard visualization tools only show counts of readings, and do not allow to correct the recurring technical deviation.

Finally, based on the link between the obtained scores and / or parameters and the visual data per chromosome, a user or practitioner may decide to change the threshold value used to define the presence of aneuploidy. The user may therefore decide to aim for a higher sensitivity (e.g., to be less stringent about the rise / fall of the parameter or scores) or higher specificity (e.g., to be more stringent about the rise / fall of the parameter or scores) .

The platform may be provided with other features that provide a more accurate analysis of the molecular data obtained from the biological sample.

As mentioned earlier, the methodology and the platform allow a certain degree of freedom to the user. In addition to defining own thresholds and thresholds, the user can also define own reference sets of genomes to be used to calculate the scores and / or other information such as the fetal fraction, gender determination, etc. using its own reference set, a user can better detect the recurring technical variation of the lab (different wet lab reagents and protocol, different NGS instrument and platform, different operator, different ...) and therefore have a better suitable reference data set for the determined lab. In order to ensure the robustness of a new reference series, methodologies are provided for the removal of outliers from the reference series. E.g. if 100 reference samples are used in a reference set, there will be 100 reference sets for every 50 kb bin. If a predetermined percentage of outliers is removed (e.g., 5% of the results), the reference set can be made more robust for variation in the reference set.

In one embodiment, a method is provided for performing CNV calling. The term "CNV calling" means a methodology that determines the limits of a CNV or segmental aneuploidy. The mentioned limits must be interpreted as the approximate chromosomal coordinates.

These limits are later used for cross-reference with CNV reference genomic databases containing previously observed CNVs (optionally annotated, e.g., benign or pathogenic). CNV calling can be performed by various methodologies. Some of them were developed in the array CGH field (where one or adjacent series of eCGH probe (s) can be considered as the equivalent of a bin), others are more specific to NGS data.

In another embodiment, the said platform allows CNV quantification. By the term CNV quantification is meant the determination of the absolute number of copies (or the expected range) of the observed CNV. The latter allows the determination of whether a CNV comes from the mother (very high value) or rather from the fetus (very low value). The aforementioned CNV quantification is preferably done after CNV calling. In another, more preferred embodiment, said CNV calling takes into account knowledge on the cell-free fraction.

In one embodiment, the said platform allows CNV signature recognition. The term "CNV signature recognition" refers to a methodology for determining whether a specific combination of CNVs (and the amount thereof) is present in a sample. CNV signature recognition is preferably performed after CNV calling and CNV quantification.

All of the above methodologies use one or more CNV reference genome databases that contain known CNVs. The above methodologies may be based on the alignment of sequences obtained from a biological sample against said one or more CNV reference databases, or else aligning the sequences obtained from a biological sample against a reference genome and later identifying the readings aligned with specific areas of interest (ie the areas identified as CNVs in the CNV reference databases). Linking with such reference databases allows the identification of CNVs that are (probably) pathogenic and therefore the (clinical) accuracy is increased: if a CNV is observed, and it contains a (likely) pathogenic area, it could be very relevant and further require patient follow-up.

The platform according to the present invention allows a high degree of freedom for the user or professional. Said platform will preferably be compatible with various data formats such as fastq, fastq.gz, call and bam files. In a further embodiment, said platform may comprise data formats that are directly streamed from the sequencing platform being used (e.g., NGS instrument). The latter greatly reduces the waiting time for data upload since the upload takes place simultaneously with the sequencing reaction. The total time-to-result will therefore be optimized, which is advantageous for the user.

In one embodiment, said platform is compatible with sequencing data from various sources, including SNRT (single-molecule real-time) sequencing data. Algorithms may be present that allow processing of SMRT data (such as, for example, PacBio) and derive epigenetic information (e.g., methylation or other modifications) from said SMRT data. The latter allows re-identification of parameters that indicate the detection of aneuploidy or could assist in determining the fetal fraction, or determining quality metrics.

The platform of the present invention is inherently compatible with many different types of NGS library preparation kits and protocols and NGS sequencing platform. This is an advantage since a user will not have to invest in special NGS sequencing platform or NGS library preparation kits specific to a specific application, but the user can use the preferred platform and the preferred kit instead. Moreover, it offers a user a certain degree of flexibility with regard to the material to be used. If newer or cheaper instruments or kits become available, a user can easily change.

As mentioned above, the present methodology is compatible with cell-free DNA extracted from various types of biological samples, including blood, saliva, blastocoel fluid, and urine. The use of urine or saliva instead of blood would provide a truly non-invasive sample type and allow, for example, testing at home and dispatch of the sample to the test lab. This is clearly an additional benefit compared to other methods of obtaining samples such as blood collection.

Examples

Preparation and seauencina of the sample

1. Blood collection, plasma separation and cell-free DNA extraction One test tube (10 ml) of maternal blood is collected in Streck test tubes and stored at 4 ° C. The blood is collected via a standard phlebotomy procedure.

The plasma (+/- 5 ml) is separated for a maximum of 72 hours after the blood has been collected by the standard double centrifugation method: • The blood sample is centrifuged at 2000xg for 20 minutes (this can take place at room temperature), without the use of the brake . The plasma is then transferred to either three 1.5 ml low binding test tubes or a single 5 ml low binding test tube. A second centrifugation at 13000xg takes place for 2 minutes (this can take place at room temperature). • The plasma is transferred to sterile 1.5 ml or 5 ml low binding test tubes for storage at -20 ° C prior to cell-free DNA (cfDNA) extraction.

The pale yellow coating can optionally be saved for later testing. Genomic DNA from the mother of the pale yellow coating can be examined to confirm or rule out abnormalities of the mother.

The cell-free DNA is extracted from the plasma using the QIAamp Circulating Nucleic Acid Kit (Qiagen) according to the manufacturer's recommendations, with a final elution volume of 60 pi. The DNA samples are stored at -20 ° C if they are not immediately used for library preparation. 2. cfDNA quantifiers

The extracted cfDNA is quantified using a Qubit fluorometer. The concentration of the cell-free DNA is usually 0.1-1 ng / µl. 3. Library preparation 25 μΙ of the extracted cfDNA is used as starting material for library preparation.

During library preparation, the DNA samples are adapted for next-generation sequencing. Adapters are added to the ends of the DNA fragments.

The sequencing libraries are prepared using the TruSeq ChlP library preparation kit (Illumina) with certain modifications to the manufacturer's protocol by lowering the reagent volumes to allow the generation of sequencing libraries using low starting amounts of DNA.

The library preparation protocol can be summarized as follows: (Note: The grain-based size selection for removing large DNA fragments and removing small DNA fragments described in the protocol is NOT used.)

Final repair of the DNA fractions: 1. Add 5 μΙ Resuspension buffer and 20 μΙ final repair mix to the 25 μΙ starting material (total = 50 μΙ)

The grain-based size selection for removing large DNA fragments and removing small DNA fragments described in the protocol is NOT used.

2. Incubate for 30 minutes at 30 ° C. 3. Add 80 μΙ undiluted AMPure beads to the 50 μΙ sample mixture after final repair. 4. Wash the pellets twice with 190 μΙ 80% EtOH. 5. Re-suspend the dried pellet with 10 μΙ resuspension buffer.

Transfer 9 μΙ of the supernatant layer to a new test tube.

Adenvlation of the 3 'ends 1. Add 6.25 μΙ A-tail mix

2. Heat for 30 minutes at 37 ° C + 5 minutes at 70 ° C

Liaation of the indexed horse-tail end adapters with the DNA 1. Adapters are diluted 1 / 2nd with resuspension buffer => add 2.5 μΙ to the sample

2. Add 1.25 μΙ ligation mix (no resuspension buffer). Incubate for 30 minutes at 30 ° C 3. Add 2.5 μΙ stop ligation buffer 4. Add 21 μΙ AMPure for cleaning 5. Wash the beads twice with 190 μΙ 80% EtOH. 6. Re-suspend the dried pellet in 27 μΙ resuspension buffer. Transfer 25 μΙ of the supernatant layer to a new test tube. 7. Add 25 μΙ AMPure for cleaning. 8. Wash the pellets twice with 190 μΙ 80% EtOH. 9. Re-suspend the dried pellet in 12.5 μΙ resuspension buffer. Transfer 10 μΙ of the supernatant layer to a new test tube.

Enriching DNA fractions 1. Prepare the PCR mix by mixing 2.5 μΙ PCR Primer Cocktail and 12.5 μΙ PCR Master Mix for each sample. 2. PCR conditions: 98 ° C for 30 seconds 15 cycles of:

98 ° C for 10 seconds 60 ° C for 30 seconds 72 ° C for 30 seconds Keep 72 ° C for 5 seconds at 4 ° C 3. Add 25 μΙ AMPure for cleaning 4. Wash the pellets twice with 190 μΙ 80% EtOH . 5. Re-suspend the dried pellet in 32.5 μΙ resuspension buffer. Transfer 30 μΙ of the supernatant layer to a new test tube. 6. Use 2 μΙ of the sample for Qubit quantification and 2 μΙ for fragmenta after lysis (see next section). 4. Quality control of library

Good cell-free DNA isolation and NGS library preparation are tested by analyzing each library on the fragment analyzer (Advanced Analytical Technologies Ine., Germany) prior to sequencing, to assess: • size distribution (confirm appropriate size profile using concentration, peak ratio, peak height, ...), • the quality of the library. High molecular weight fragments containing samples will be classified as samples that are eligible for sequencing (indicates contamination with genomic DNA from the mother).

Typical libraries exhibit a narrow size distribution with a peak at about 300-350 bp.

In addition, a Qubit quantization step is performed so that the enrichment reaction will take place with the appropriate amount of inserted DNA material. The concentration of DNA is usually 15-30 ng / µl. 5. Normalize and group libraries

Samples are indexed during library preparation and up to 24 samples are normalized and grouped in equal volumes for multiplex sequencing across both lanes of an Illumina HiSeq2500 flow cell. 6. NGS run

Sequencing is performed on the HiSeq 2500 (Illumina) in Fast Run mode where 50 bp single-end readings are produced.

Detection of aneuploidy in a biological sample: validation of the methodology • Mapping and filtering of the assigned readings

The 50 bp single-end sequence readings from a test sample are assigned to the reference genome GRCh37.75 with BWA backtrack. With Picard tools, duplicated readings are removed and based on mapping quality, readings that refer to multiple locations are ignored. Readings that refer sub-optimally to multiple locations are also deleted. To reduce the variability of samples, we only use the readings that perfectly match the reference genome (i.e. no mismatches and no openings are allowed). Finally, readings that are blacklisted in an internally compiled list of areas are also deleted. These blacklisted areas include common polymorphic CNVs, collapsed repeats, DAC blacklisted areas generated for the ENCODE project, and the undefined part of the reference genome (i.e., the eNs). * Calculation of genomic representation

The reference genome is subdivided into 50 kb pieces and the number of readings from the test sample is counted per piece. These counts of readings are corrected according to the GC levels of the pieces with locally weighted scatter plot smoothing (Loess regression). These GC-corrected reading counts are then divided by the total sum of all autosomal GC-corrected reading counts and multiplied by 107. This is defined as the genomic representations (GR) per piece. A sliding window is applied to these per-piece GR values and the sum of these GR values is determined for all consecutive 100 pieces. The windows are each shifted in time by 1 piece (i.e., 50 kb). In this way a GR value is obtained per window of 5 Mb. Similarly, for each autosome, the sum of the piece GR values is calculated to obtain a GR value for each autosome in the test sample. * Comparison with a reference series

In a reference set of 100 normal samples (50 male and 50 female pregnancies), the GR values are calculated for all autosomes and for all 50 Mb windows as described above. For each autosome and window of 5 Mb, the mean μ and the standard deviation σ of the GR scores are calculated over all 100 reference samples. In this way, a Z score can be calculated for each window and every autosome i in a test sample, defined as

where GRt is the GR value in the test sample for window or autosome i and μ £, σ; the average and standard deviation, respectively, of the GR scores measured in the 100 window or autosome i reference samples.

Based on the 22 Z-scores of the car steams in a test sample, a ZZ2 score was calculated for each autosome as

wherein the Z-score Zt of autosome i in the test sample is compared with the median and the standard deviation (sd) of the 22 Z-scores obtained for all 22 autosomes in the test sample.

Alternatively, a ZofZ score is calculated as

wherein the Z score Z {of chromosome i in the test sample is compared with the median and median absolute deviation (mad) of the 22 Z scores obtained for all 22 autosomes in the test sample. The ZZ2 and ZofZ scores quantify the deviation of the Z score of the target autosome from? all Z scores observed in the test sample. This robust version of the Z-or-Z scores makes no assumptions about the aneuploidy status of the autosome and the other autosomes.

Based on the Z scores calculated for all 5 Mb windows in the test sample, the BM score of each car steam / is calculated as the median of the Z scores for all windows in the target autosome:

where the median of the Z-scores is calculated over all windows j on autosome / '.

This BM score reflects the magnitude of the deviation: aneuploidies will result in higher BM values, while smaller, segmental CNVs will have less influence on the median of the Z scores and result in lower BM scores.

To distinguish between deviation-related BM scores and increased BM values due to noise in the data series, the OM score for an autosome is calculated as the median of the Z scores of all windows of the other autosomes:

the median being calculated over all, absolute Z-scores for 5 Mb windows j that are not on autosome / '.

Finally, a quality score (QS) is calculated for each test sample

with; over all autosomes except for the 2 autosomes with the highest and lowest Z score. This score will identify test samples with poor quality that result in unreliable aneuploidy calling. A greatly increased QS score can also refer to DNA samples containing at least a fraction of DNA from a tumor.

A threshold value can be defined for each of the parameters calculated above. A threshold value of 2, 2.5 or 3 can be chosen based on standard statistical considerations. In the context of the Z score, this means that the probability that the test result is normal (i.e. the obtained GR score is similar to the GR scores for the same area in the reference series) is very small. To make a test more specific, the threshold value could be increased. To make a test more sensitive, the threshold value could be lowered. These threshold values can be determined for each of the parameters, and can differ for each of the parameters. For example, it is conceivable that threshold values for BM and OM are set to 1 while they are set to 3 for the Z score and ZZ score. Negative thresholds can also be used.

Example 1

1.1 ZZ2 calculation for chromosome 21 for sample A

For sample A, the median of the Z scores for all chromosomes is -0.1641 and the standard deviation is 2.494. For chromosome 21 we have a Z score of 11,147 and this results in a ZZ2 score of 4,5347, above the threshold of 3. (Figure 1)

This automated classification as an abnormal chromosome based on ZZ2> 3 has also been confirmed by visual inspection of the graph. This sample was tested with an invasive test and the trisomy was confirmed.

None of the other chromosomes in sample A have elevated ZZ2 scores (Figure 2).

The graphs of the Z scores of these other chromosomes are also not indicative of aneuploidy. On chromosome 11, for example, we have a ZZ2 score of 0.319 and this results in the graph as shown in Figure 3.

The graph in Figure 3 shows the BM scores of all autosomes. The latter confirms that chromosome 21 is highly abnormal, while other potential autosomal abnormalities will cover less than half the chromosome.

1.2 ZofZ calculation for chromosome 21 for sample A

For sample A, the median of the Z scores for all chromosomes is -0.164 and the median absolute deviation is 0.819. For chromosome 21 we have a Z score of 11,147 and this results in a ZofZ score of 13.817, far above the threshold of 3. If we plot the Z scores in a graph, this clearly indicates trisomy 21 (Figure 9) . This sample was tested with an invasive test and the trisomy was confirmed.

Based on these ZofZ scores and a threshold of 3, none of the other chromosomes would be called aneuploidy (Figure 10).

The graphs of the Z scores of these other chromosomes are also not indicative of aneuploidy. For example, on chromosome 11 we have a ZofZ score of 0.973 and this results in the graph as shown in Figure 11.

1.3 ZZ2 calculation for chromosome 16 for sample B

In sample B, the median of the Z scores for all chromosomes is -0.2651 and the standard deviation is 1.464. For chromosome 16 we have a Z score of 5.754 which results in a ZZ2 score of 4.111, which is higher than the threshold of 3.

Namely, if the Z scores are plotted in a graph (Figure 5), this clearly indicates a trisomy 16. This sample was tested with an invasive test and the trisomy was confirmed.

None of the other chromosomes in sample B have elevated ZZ2 scores (Figure 6).

The graphs of the Z scores of these other chromosomes are also not indicative of aneuploidy. For example, on chromosome 1 we have a ZZ2 score of -0.459 which results in the graph as shown in Figure 7.

The graph of BM scores of all autosomes as shown in Figure 8 confirms that chromosome 16 is highly abnormal, while other, possibly abnormal autosomal abnormalities will cover less than half of the chromosome.

1.4 ZofZ calculation for chromosome 16 for sample B

For sample B, the median of the Z scores for all chromosomes is -0.265 and the median absolute deviation is 0.685. For chromosome 16 we have a Z score of 5.754, which results in a ZofZ score of 8.772, above the threshold of 3. If we plot the Z scores in a graph, this clearly indicates trisomy 16 (Figure 12). This sample was tested with an invasive test and the trisomy was confirmed.

Based on these ZofZ scores and a threshold of 3, none of the other chromosomes would be called aneuploidy (Figure 13).

The graphs of these other chromosomes are also not indicative of aneuploidy. For example, on chromosome 1 we have a ZofZ score of -0.980 (Figure 14).

1.5 Assessment of sample quality via ON and QS

Samples A and B were two clear cases in which one chromosome is trisome, while all other autosomes behave diploid. The following graphs show the OM values of the autosomes in samples A and B and all OM values confirm that this was a successful experiment. This is also confirmed by low QS scores (i.e., 0.576 for sample A and 0.652 for sample B, see Figures 15 and 16).

Example 2

In sample C we observe a ZofZ score of 4.141 for chromosome 7, while the BM score remains low. Figure 17 shows a graph of chromosome 7, where a portion of the chromosome that is likely to have a higher copy number is observed. This may indicate the mother's CNV.

ZofZ scores can therefore also be indicative of segmental CNVs.

Automated classification of the chromosome using ZofZ> 3 is very sensitive to inclusion of CNVs (see example C) or larger aneuploidies (see samples A and B). ZofZ can therefore be used to identify abnormal chromosomes, and visual inspection of the graph can confirm the presence of such abnormalities.

Automated classification of the chromosome using a combination of ZofZ> 3 with another parameter, can further improve the specificity of the automated classification, and add more granularity to the results. If, for example, ZofZ> 3 and BM <1, this may indicate the presence of a CNV (see sample C), while if, for example, ZofZ> 3 and BM> 1, this may indicate the presence of greater aneuploidy (see samples A and B).

Example 3: ZofZ more sensitive than ZZ

The ZofZ score can be more sensitive compared to ZZ for the identification of CNVs. ZofZ score can also be more sensitive compared to ZZ for the identification of CNVs or larger aneuploidies in noise samples. 3.1 Segmental abnormality on chromosome 21

In Sample D, a ZofZ score of -6.873 for chromosome 21 was observed, while the ZZ score as described in Bayindir et al. 2015 (which was used as a reference method) resulted in a ZZ score of -2.341 (ie not significant) . In the graph of that chromosome as shown in Figure 18, we observe a part of the chromosome that appears to be underrepresented. This may indicate CNV from the mother or fetus. 3.2 Segmental abnormality on chromosome 15

Based on the ZZ score and the decision tree described in Bayindir et al. 2015, this chromosome would be classified as a normal profile (ZZ = -2.730). However, the ZofZ score of -4.51 draws attention to this chromosome (see Figure 19). The BM score is -0.6, indicating that the deviation is partial, as could also be deduced from the visual inspection of the graph. 3.3 Segmental abnormality on chromosome 22

In sample E, a ZofZ score of 3.029 for chromosome 22 was observed (see Figure 20), while the ZZ score as described in Bayindir et al. 2015 resulted in a ZZ score of -0.629 (i.e., not significant).

Based on the ZZ score and the decision tree described in Bayindir et al. 2015, this chromosome would not be designated and classified as normal. This would be due to the fact that this sample also has a trisomy 18 (see Figure 21). 3.4 Segmental abnormality on chromosome 20

In sample I, the ZZ score as described in Bayindir et al. 2015 would be lower than 3 (i.e., ZZ = 2,195), while the ZofZ score is equal to 3.31. The BM score is also equal to 1.51, which indicates that more than half of the Z scores have risen. The OM score of 1.05 shows that this data set contains noise earlier (see Figure 22). 3.5 Indication of chromosome monosomy 22

In sample F we see a ZofZ score of -3.094 for chromosome 22, while the ZZ score calculated as described in Bayindir et al. 2015 equals -1.771 (i.e. not significant) (Figure 23). The ZZ score would result in a normal value. Sample F has a trisomy 21 that may mask this monosome behavior of chromosome 22. Note that the trisomy 21 was confirmed via invasive follow-up. There were no follow-up data for chromosome 22.

The latest results all show that the method of the present invention is more sensitive than the methodologies well known in the art. This visual representation of the data as shown in the figures allows automated classification and interpretation. Moreover, the visualization according to the present invention allows a distinction to be made between technical noise and noise due to aneuploidies.

If the data contains noise, and varies along the Z = 0 axis (i.e., both higher and lower than 0), the chance of technical noise is greater. While if the data does not vary along the Z = 0 axis for a large chromosomal segment, the chance of renal aneuploidy is greater. The graph on figure 24 clearly shows that the view is not due to technical noise, but to a different situation.

Example 4: Gender determination

The gender of the samples can be determined by judging the number of readings referring to twenty 50 kb pieces on chromosome Y that were empirically selected to be specific for male pregnancies (see Bayindi 2015 for more details). In case at least 3 or more Y-specific pieces contained more than 1 reading, the gender was determined to be male. In case a maximum of 1 piece contained 1 reading or none of the 20 pieces contained a reading, it was determined that the gender was female. In all other cases, no gender determination was made and the gender was said to be undetermined. For example, this could be due to a vanishing twin when the blood sample contains fetal DNA from two fetuses instead of one.

In a series of 249 successful experiments (ie QS score was lower than 1.5 and the number of readings remaining after all filtering steps was higher than 7,000,000), gender was determined and this resulted in a series of 116 (46, 59%) female pregnancies, 131 (52.61%) male pregnancies and 2 (0.80%) indefinite pregnancies. The graphs as shown in Figures 24 and 25 for these 249 samples show the number of readings assigned to the Y chromosome (after filtering the BAM file). 4.1 Fetal fraction determination for male pregnancies

Once the gender has been determined, the fetal fraction can be determined for male pregnancies. For male pregnancies one can take advantage of the fact that the fetal DNA will only have 1 copy of X and a copy of Y, instead of the 2 copies of X that are present in the DNA of the mother. This makes it possible to determine the fetal fraction in two ways. On the basis of the X chromosomes, the fetal fraction can be determined as twice the difference at the 50 kb piece level between the median number of readings mapped against the autosomes and the median number of readings mapped against chromosome X, divided by the median number of readings mapped against the autosomes. This can be written as the following formula:

Secondly, the fetal fraction can also be estimated based on the Y chromosome since all readings relating to chromosome Y must in theory come from the fetal DNA. The fetal fraction based on chromosome Y is defined as twice the median number of GC-corrected readings mapped compared to Y over the median number of GC-corrected readings mapped to the autosomes, or in a formula:

A disadvantage of this approach is that the fetal fraction can only be determined for about half of the samples (i.e. only for male pregnancies). In our previous examples, fetal fractions were calculated. The results are shown in Table I.

Figure 25 shows the X and Y-based fetal fractions for the 131 male pregnancies identified in Example 4.1.

Example 6: CNV-based approach for determining a minority fraction fbiiv. fetal fraction!

The methodology of the present invention can be illustrated by starting from a sample with a fetal fraction of 10%, which was sequenced with a coverage of 0.1X to give 50 bp readings. Furthermore, a CNV of 10 Mb must be assumed.

In the case of a normal sample (i.e., 2 copies of the CNV) consisting of 100% DNA from the patient, 20,000 readings referring to that CNV region are expected. In case the patient has only 1 copy of the particular CNV, 10,000 readings are expected in that area; in the case of 0 copies, no readings are expected.

In the case of a minority fraction of 10%, the following cases can be expected (Table II): # copies # copies fetus Expected # female readings

Table II

If it is found that the CNV has a coverage that is far too low and is not zero, it can be concluded that the mother does not have this area (0 copies) and that the readings observed are from the minority DNA. Consequently, the CNV can be considered as informative for the determination of the fetal fraction.

Therefore, the minority fraction can be determined as follows: 2 * observed readings / expected readings = 2 * 1,000 / 20,000 = 10%

As an estimate for the expected number of fragments, the global coverage of the sample can be used as well as the length of the CNV. A correction of the general reading depth can be made and the expected coverage based on all samples can be determined, eg by taking the average after excluding the upper and lower 10% of the samples or using the median coverage.

To correct for recurring technical noise, a CNV-specific attribute can be calculated using samples with known fetal fraction, whereby the attribute can correct the obtained reading counts for recurring technical noise on that particular CNV.

The sequences can be obtained as described above, except that filtering of readings in the areas on the blacklist was not used.

Calculation of coverage per CNV

A CNV reference genome database was used, containing 581774 CNVs. Note that the CNVs in this series were not all unique CNVs since the database contains overlaps.

The readings were aligned against the reference genome. For each CNV, the number of readings that overlapped at least X bases with the CNV regions was counted, where X can be any value between 1 and 50. In the current example, X was set to 1. This results in a matrix with the raw counts. Unprocessed counts equal to or lower than X are defined as being equal to 0 since they have minimal overlap with the CNV.

Because the raw counts were not corrected for the total number of readings, the number of filtered readings was extracted for each sample and it was corrected by the following equation: normalized count = 10,000,000 * raw count / total number of readings; 10,000,000 was chosen at random and could be replaced by any other value.

Previously determined fetal fractions obtained for a series of male pregnancies were imported. These fetal fractions are based on the counts on the X and Y chromosome and use the fact that, in male pregnancies, the fetal DNA has only 1 copy of chromosome X and one copy of chromosome Y; while the DNA of the mother has two copies of chromosome X. Consequently, the X- and Y-based method of the fetal fraction can only provide an estimate of the fetal fraction for male pregnancies (as opposed to CNV-based method of the fetal fraction presented here).

The previously obtained fetal fractions were filtered to remove samples with a fetal fraction smaller than Y, Y can assume any value between 0.05 and 1. In the present example, said Y was set to 0.3, since this was considered to be the maximum fetal fraction that one could routinely encounter.

Determining informative CNVs in a series of CNVs

Information CNVs were determined in a next step. For each CNV, samples were identified for which only the fetus had 1 or 2 copies and where the mother had no copies. All counts are therefore derived from the fetus. It was then checked whether these fetal counts for male pregnancies correlate with the X / Y-based fetal fractions, resulting in an X-based and Y-based correlation. This provided insight into whether the CNVs were informative for fetal fractions.

As an example, the following 3 random CNVs were considered, covering different lengths: 1. Chr 1: 72,773,259-72,798,581; this is an area of 25kb (25.323bp) 2. Chr 1: 148.539.255-149.765.886; this is an area of 1 Mb (1,226.632bp) 3. Chr 9: 38.725.590-71.025.693; this is an area of 32Mb (32,300,104bp)

In a scatter diagram (not shown) of the normalized counts (normalized with respect to the total number of readings) for CNV 1, 3 peaks are observed (which are likely to correspond to 0, 1 and 2 copies of the mother). For CNV 2, 4 large peaks were observed (which probably correspond to 0, 1, 2 and 3 copies of the mother). No peaks could be observed in the third and largest CNV. The analysis was therefore continued with CNV 1 and 2.

To find local minima, a density function was applied and the signs of the derivative of the density functions were checked.

The two local minimums were at: 1. CNV 1: 18.39 and 67.96 2. CNV 2:

766.69; 1160.26; 1538.69; 1568.97; 1705.20 and 1856.57 (see figure 29).

The normalized counts were extracted for those samples that had normalized counts that were smaller than the smallest local minimum and greater than 0. It was assumed that the counts for these samples were derived primarily from the fetal DNA and made minimal to low DNA contribution from the mother.

Based on the normalized counts, the expected count was calculated for each CNV. The normalized counts were normalized to a randomly chosen value of 10,000,000 readings.

With this expected count, an estimate of the fetal fractions was calculated as: 2 x 100 x observed counts / expected counts.

This estimate of the fetal fraction is correlated with the actual fetal fraction. However, it was found that it is not identical. The estimated fetal fraction can in fact be seen as follows: estimated fetal fraction = factor * actual fetal fraction.

This factor is a constant factor that can be considered as an attribute of each CNV in the data series, and which must be determined empirically (see below).

Example of determining the fetal fraction: a) for CNV 1:

Fetal fraction can be estimated as 2 * 100 * normalized count / expected counts = 2 * 100 * normalized count / 88.58

Figure 28 shows the calculated fetal fractions for all male pregnancies with fewer than 18.54 counts for CNV 1 (i.e., 142 samples) versus the in-house fetal fraction calculated via chromosome X (left) or chromosome Y (right). Correlations are equal to 0.54 and 0.56 with the X and Y-based fetal fractions, respectively. b) for CNV 2

Fetal fraction can be estimated as: 2 * 100 * normalized count / expected counts = 2 * 100 * normalized counts / 4,290

Figure 29 shows the calculated fetal fractions for all male pregnancies with fewer than 359.20 counts for CNV 2 (i.e., 17 samples) versus the fetal fraction calculated via chromosome X (left) or chromosome Y (right). Correlations are equal to 0.601 and 0.963.

For all CNVs, the local minimum, the local maximum, the number of peaks, the number of male pregnancies with a count lower than the smallest local minimum and the correlations with the X / Y-based fetal fractions were calculated.

Autosomal "pseudo-informative" CNVs were selected based on this information. • X or Y-based correlation greater than A (between 0.01 and 1, A was set as 0.5 in this example) • based on at least B (between 1 and 100 or more, B was set to 8 in this example) male pregnancies • with more than C peaks (between 0 and 5, C was set to 2). • exclude the CNVs on the X and Y chromosome. • compare the first local minimum with the third local maximum. The ratio between the third local maximum and the first local minimum must be greater than D (between 0.1 and 100, D is set to 3).

This can greatly reduce the number of CNVs to anywhere between 1 and 100,000 or more "pseudo-informative" CNVs. In this example, approximately 5,000 "pseudo-informative" CNVs were identified.

Many overlapping CNVs were identified within the pseudo-informative CNVs list. The list was therefore cleaned up by one or more of the following methodologies: • Per set of overlapping CIWs, only A (A = between 1 and 100 or more, where A was set as 1 in this example), ie the one with the highest average correlation (ie average of the X and Y based correlations. • Removing duplicates • For each similar CNV, only the longest CNV was used.

Note that cleaning is optional, but can greatly reduce the number of CNVs to anywhere between 1 and 100,000 or more "informative" CNVs. In this example, the number was reduced to around 100 "informative" CNVs.

In a later step and for each of the informative CNVs, the normalized counts were scaled to the X / Y-based fetal fractions. Therefore, for each CNV, it was evaluated how the reading counts predict the fetal fraction. The method is limited to CNVs with a correlation higher than D (where D is between 0.01 and 1 and is set as 0.5 in this example). In case the X-based correlation was higher than D, while the Y-based correlation was lower than D, only the X-based fetal fractions were taken into account. Note that the reverse is also possible.

CNV 1: 1,541,063-1,541,536 was considered as an example. A regression line of the X (Y) -based fetal fractions versus the normalized counts was placed before the small counts seen in male pregnancies from successful experiments. This resulted in 2 regression lines, each with a cut-off section and direction coefficient, for all CNVs or all pseudo-informative CNVs, or all informative CNVs.

With the direction coefficient and the cut-off part of these regression lines, the corresponding, small normalized counts were scaled for each CNV. Note that this scaling can occur for all samples (i.e., male and female samples). The following graphs show the scaled counts versus the X / Y-based fetal fractions for male pregnancies who had small counts for CNV 1: 1,541,063-1,541,536.

For each CNV and for all samples with small counts for that particular CNV, two estimates of the fetal fraction were obtained, the X and Y-based fetal fraction. The third estimate was the average of the X and Y-based scaled counts.

In case one of the 2 correlations, e.g. the X-based correlation, is lower than E (where E is between 0.01 and 1, and set as 0.5 in this example, E = 0.5 in this example), then the average is equal to the Y-based scaled counts.

For each sample, three estimates for the fetal fraction for all CNVs showing a small number of readings within the sample were obtained: - X-based estimate - Y-based estimate - average-scaled estimate

In a final step, the average of the CNV-based fetal fraction over all CNVs was obtained. To check whether this average represents the chromosome X and Y-based fetal fractions for male pregnancies, they are plotted versus each other for the successful experiments.

It is clear from Figure 31 that there is a clear correlation between the CNV-based determination of the fetal fraction and the X and Y-based determination of the fetal fraction. To use the CNV-based method of the fetal fraction to determine the actual fetal fraction, the scaling can be empirically adjusted (i.e., the regression line can be adjusted to have a direction coefficient of 1).

Claims (30)

CONCLUSIONS A method for determining the presence or absence of a fetal chromosomal aneuploidy in a pregnant woman, the method comprising: - providing the sequences of at least a portion of the nucleic acid molecules contained in a biological sample that is obtained from said pregnant woman, wherein said biological sample contains cell-free DNA from both the mother and the fetus; - aligning said obtained sequences with a reference genome; - counting the number of readings on a series of chromosomal segments and / or chromosomes whereby reading counts are obtained; - normalizing said reading counts or a derivative thereof to a normalized number of readings; - obtaining a first score of said normalized readings and obtaining a set of scores of said normalized readings, wherein said first score is derived from said normalized readings for a target chromosome or chromosomal segment, and wherein said set of scores is a set of scores derived from the normalized number of readings for a set of chromosomes or chromosome segments containing the chromosomal target segment or chromosome; - calculating a parameter p on the basis of said first score and said set of scores, wherein said parameter represents a ratio between * said first score, corrected by a summary statistic of said set of scores, and * a summary statistic of the said set of scores; and - comparing said parameter p with a threshold value, said threshold value being a requirement for the presence or absence of aneuploidy of the target chromosome or chromosomal segment. The method according to claim 1, characterized in that said number of readings is recalibrated to correct for GC content and / or total number of readings obtained from said sample. The method according to claim 1 or 2, characterized in that said normalization occurs via comparison with data obtained from corresponding chromosomal segments or chromosomes from a reference series. The method according to any one of claims 1 to 3, characterized in that said summary statistic is the average, the median, the standard deviation, the average absolute deviation or the median absolute deviation. The method of any one of claims 1 to 4, wherein the sequencing is randomly performed on a portion of the nucleic acid molecules contained in the biological sample. The method of any one of the preceding claims, wherein the biological sample is maternal blood, plasma, serum, urine, blastocoel fluid, transcervical fluid, or saliva. The method according to any of the preceding claims, wherein the chromosomal target segment is selected from Table 1, and / or from a part or a window derived from chromosome X, Y, 6, 7, 8, 13, 14, 15, 16, 18, 21 and / or 22. The method according to any of the preceding claims 1 to 6, characterized in that said target chromosome is selected from chromosome X, Y, 6, 7, 8, 13, 14, 15, 16, 18, 21 and / or 22. The method of any one of the preceding claims, wherein said threshold value is determined using standard statistical considerations, or empirically determined using biological samples. The method of any one of the preceding claims, wherein said score is calculated as: wherein i is a chromosome or chromosomal segment or the target chromosome or chromosomal target segment. The method according to claim 10, characterized in that said parameter p is calculated as: wherein (Zj) represents a set of scores derived from chromosomes or chromosomal segments i, a, b, ... where i corresponds to the chromosomal target segment or chromosome. The method according to any of the preceding claims, comprising the calculation of secondary parameters, wherein said secondary parameters are a requirement for the presence of said aneuploidy and / or a measure of the quality of the sample. The method of claim 10, wherein said secondary parameters are compared to a threshold value. The method according to any of claims 12 or 13, wherein said presence or absence of an aneuploidy is determined by comparing said parameter with a threshold value and comparing one or more secondary parameters with corresponding threshold values. The method of any one of the preceding claims, wherein the fetal fraction of the sample is determined. The method of claim 15, wherein said determining the fetal fraction comprises the steps of: - counting the number of sequences that align with a predefined set of polymorphisms; - comparing the obtained number of sequences with the expected number of sequences to identify the informative polymorphic site (s) for the sample; - calculating on the basis of the obtained number of sequences for said informative polymorphic site (s) an amount, said amount being an indication of the fetal fraction. The method of claim 16, wherein said amount is calculated using linear scaling based on informational polymorphism-specific attributes. The method of any one of claims 15 to 17, wherein said amount indicative of the fetal fraction serves as a quality check of said sample. A computer program product comprising a computer-readable medium encoded with multiple instructions for controlling a computer system for performing an operation for performing prenatal diagnosis of a fetal chromosomal aneuploidy in a biological sample obtained from a pregnant female patient wherein the biological sample contains nucleic acid molecules, the processing comprising the steps of - receiving the sequences of at least a portion of the nucleic acid molecules contained in a biological sample obtained from said pregnant woman, said biological sample comprises cell-free DNA from both the mother and the fetus; - aligning said obtained sequences with a reference genome; - counting the number of readings on a series of chromosomal segments and / or chromosomes, whereby counts of readings are obtained; - normalizing said number of readings or a derivative thereof to a normalized number of readings; - obtaining a first score of said normalized readings and a set of scores of said normalized readings, wherein said first score is derived from the normalized readings for a target chromosome or chromosomal segment, and wherein said set of scores is a set is scores derived from the normalized readings for a series of chromosomes or chromosomal segments containing the target chromosomal segment or chromosome; - calculating a parameter p on the basis of said first score and said set of scores, wherein said parameter represents a ratio between * said first score, corrected by a summary statistic of said set of scores, and * a summary statistic of the said set of scores; and - comparing said parameter p with a threshold value, said threshold value being a requirement for the presence or absence of aneuploidy of the target chromosome or chromosomal segment. The computer program product of claim 19, further comprising operations for calculating one or more secondary parameters, said secondary parameters being a requirement for the presence of said aneuploidy and / or a measure of the quality of the sample. A computer program product according to any one of the preceding claims, comprising operations for determining the fetal fraction. A computer program product according to any one of the preceding claims, further comprising operations for performing CNV calling, CNV quantification and / or CNV signature recognition. A kit comprising a computer program product according to any of claims 19 to 22 and a protocol for obtaining the sequences of at least a portion of the nucleic acid molecules contained in a biological sample obtained from a pregnant woman, said biological sample being cell-free DNA from both the mother and the fetus. The kit of claim 23, further comprising reagents and means for obtaining said sequences. A report comprising an estimate of the presence or absence of a fetal chromosomal aneuploidy in a pregnant woman, said report comprising the parameter, one or more secondary parameters and a comparison with a threshold value as defined in any one of claims 1 to 18 and a visualization of the aforementioned readings per chromosome. A report according to claim 25, characterized in that said visualization shows said first score per window of a target chromosome and / or parameter p. A method for determining a fetal fraction in a biological sample obtained from a pregnant woman, said method comprising: - receiving the sequences of at least a portion of the nucleic acid molecules contained in a biological sample that is obtained from said pregnant woman; - counting the number of sequences that align with a predefined set of polymorphisms; - comparing the obtained number of sequences with the expected number of sequences to identify the informative polymorphic site (s) for the sample; - calculating on the basis of the obtained number of sequences for said informative polymorphic site (s) an amount, said amount being an indication of the fetal fraction. The method of claim 27, wherein said amount is calculated using linear scaling based on informational polymorphism-specific attributes. A method according to claim 27 or 28, characterized in that said polymorphisms are copy number variations with a size between 100 bp and 1 Mb, or between 1 kb and 1 Mb, or between 2 bp and 250 Mb. A computer program product comprising a computer-readable medium encoded with a plurality of instructions for controlling a computer system to perform an operation for determining or estimating the fetal fraction in a biological sample obtained from a pregnant female patient, the biological sample contains nucleic acid molecules, the processing comprising the steps of: - receiving the sequences of at least a portion of the nucleic acid molecules contained in a biological sample obtained from said pregnant woman; - counting the number of sequences that align with a predefined set of polymorphisms - comparing the obtained number of sequences with the expected number of sequences to identify the informative polymorphic site (s) for the sample; and - calculating on the basis of the obtained number of sequences for said informative polymorphic site (s) an amount, said amount being an indication of the fetal fraction.