Methods for non-invasive prenatal testing for amplification of mutations

Abstract

Disclosed herein are method, computer software product, and system embodiments for genotyping a fetus based on analysis of cell-free DNA (cfDNA) in a maternal plasma sample, in particular for identifying short tandem repeat sequence (STR) variants. Embodiments include receiving (i) maternal plasma cfDNA and (ii) sequencing data reads of maternal and optionally paternal genomic DNA (gDNA) from a pair of parents who are nurturing the fetus; identifying potential genomic loci for which the fetus can have an STR variant; for each of the potential genomic loci, determining a probability that the fetus has the STR variant; inferring maternal and optionally paternal STR alleles; processing cfDNA reads in a region of a gene that potentially contains an STR to identify read classes; calculating expected ratios between the read classes; and determining an allele that the fetus is most likely to inherit.

Description

Technical Field

[0001] The present disclosure relates to the field of prenatal genetic analysis.

[0002] References: Dashnow H, Lek M, Phipson B, Halman A, Sadedin S, Lonsdale A, et al. STRetch: detecting and discovering pathogenic short tandem repeat expansions. Genome Biology. 2018;19:121. Dolzhenko et al., ExpansionHunter: a sequence-graph-based tool to analyze variation in short tandem repeat regions. Bioinformatics 2019; 35(22): 4754-4756. Liautard-Haag C, Durif G, VanGoethem C, Baux D, Louis A, Cayrefourcq L, et al. Noninvasive prenatal diagnosis of genetic diseases induced by triplet repeat expansion by linked read haplotyping and Bayesian approach. Sci Rep. 2022;12:11423. Rabinowitz T, Polsky A, Golan D, Danilevsky A, Shapira G, Raff C, et al. Bayesian-based noninvasive prenatal diagnosis of single-gene disorders. Genome Res. 2019. https: / / doi.org / 10.1101 / gr.235796.118. Rabinowitz T, Deri-Rozov S, Shomron N. Improved noninvasive fetalvariant calling using standardized benchmarking approaches. Computational and Structural Biotechnology Journal. 2021;19:509–17. US2021 / 0340601. Yu SCY, Jiang P, Peng W, Cheng SH, Cheung YTT, Tse OYO, et al. Single-molecule sequencing reveals a large population of long cell-free DNAmolecules in maternal plasma. Proceedings of the National Academy of Sciences. 2021;118:e2114937118. Background Technology In recent years, an increasing number of gene mutations have been clinically detected through non-invasive prenatal testing (NIPT), enabling pregnant women to identify genetic abnormalities without the risk of invasive surgery.

[0003] Rabinowitz et al. (2019) described a genome-wide NIPT method called non-invasive prenatal variant detection. Using... Hoobari (The first non-invasive fetal variant detector) They are able to genotype all fetal locations (including parental loci and insertion / deletion sites) (US2021 / 0340601). Hoobari This is a tool that uses a Bayesian algorithm to predict the inheritance of single-gene diseases, regardless of their inheritance pattern or parental origin (Rabinowitz et al. 2019; US2021 / 0340601; Rabinowitz et al. 2021). This tool can estimate the probability of fetal origin by taking into account the length of each cfDNA fragment. Hoobari It can detect mutations caused by single nucleotide polymorphisms (SNPs) or small insertions and deletions (insertions and deletions of bases in the genome).

[0004] Short tandem repeat (STR) mutations occur when errors occur during DNA replication, leading to an increase in the number of repetitions of common trinucleotide sequences (such as CAG, GCG, etc.). In each disease involving this mechanism, there are two thresholds: one that distinguishes normal alleles from premutations, and a second that distinguishes premutations from whole mutations. This second threshold relates to the number of repetitions; exceeding this number of repetitions results in the phenotype being considered disease-related. These thresholds vary by disease. While invasive clinical tests for STR analysis exist (based on chorionic villus sampling or amniocentesis), no non-invasive clinical test for STR analysis has yet been proven successful. STRs can cause amplification mutations. Detecting amplification mutations is a crucial target because over 40 diseases primarily affecting the nervous system originate from the amplification of simple repetitive sequences scattered throughout the human genome, such as Fragile X syndrome.

[0005] Liaotard-Hagg et al. described an indirect method that uses linked reads to sequence and phase parental genomic sequence data (i.e., to obtain haplotype information) using markers present around amplified mutations. As a next step, the authors used... Hoobari To predict genetic haplotypes.

[0006] Yu et al. described the use of PacBio long-read sequencing to sequence long cell-free DNA molecules and infer fetal haplotypes (hereditary alleles).

[0007] Liaotard-Hagg et al. and Yu et al. both deduced the inherited alleles, but they did not predict whether amplification occurred or whether it led to a full mutation. Summary of the Invention

[0008] In its first aspect, the present invention provides a method for genotyping a fetus, the method comprising: a. Receive sequencing data reads from (i) cell-free maternal plasma DNA (cfDNA) and (ii) maternal and optionally paternal genomic DNA (gDNA) from a pair of parents that nurtured the fetus; b. Identify potential genomic loci in the fetus that may have short tandem repeat (STR) variations; c. For each of the potential genomic loci, determine the probability that the fetus has the STR variant; d. Infer the STR alleles of the mother and the arbitrarily chosen father; e. Process cfDNA reads in regions of the gene that potentially contain STRs to identify read categories, wherein the read categories include spanning reads, partial reads (including indicative and non-indicative reads), and clean reads; f. Calculate the expected ratio between the read segment categories to obtain the calculated ratio; g. Determine the most likely inherited alleles of the fetus using the following method: i. Measure the observation ratio between the read segment categories; and ii. Use the calculated ratio and the observed ratio to distinguish normal alleles from alleles with premutations or full mutations, and to distinguish alleles with premutations from alleles with full mutations; or iii. Apply Bayesian algorithm; This allows for genotyping of the fetus.

[0009] In one embodiment, the sequencing data reads are short reads.

[0010] In one embodiment, the sequencing data reads are long reads.

[0011] In one embodiment, the sequencing data reads include short reads and long reads.

[0012] In some embodiments, one or both of the gDNA sequencing data and cfDNA sequencing data are obtained by a method selected from the group consisting of: whole genome sequencing (WGS), whole exome sequencing (WES), next-generation sequencing (NGS), targeted sequencing, kit sequencing, gene sequencing, long read genome sequencing, paired-end sequencing, single-end sequencing, and amplicon sequencing.

[0013] In one embodiment, WGS or WES data is obtained through deep sequencing.

[0014] In one embodiment, the probability is determined based on at least one sequence alignment map (SAM) parameter.

[0015] In one embodiment, the probability is determined based at least on the observed template length.

[0016] In one embodiment, determining the probability includes calculating a total fetal score.

[0017] In one embodiment, the method further includes constructing a fetal size distribution and a maternal size distribution, wherein the determination of the probability in step c includes: binning the fetal size distribution and calculating a fetal score for each segment size bin, and for at least one size and at least one segment at the at least one location, calculating the probability that the segment is fetal based on the fetal score of the corresponding segment size bin to which the segment belongs.

[0018] In one embodiment, determining the probability includes applying a Bayesian procedure.

[0019] In one embodiment, the Bayesian procedure includes prior probabilities calculated using sequencing data from at least one of the parents.

[0020] In one embodiment, the method further includes recalibrating the output of the Bayesian program using machine learning.

[0021] In one embodiment, the determination of the probability is performed using fetal variation detection.

[0022] In one embodiment, determining the probability is done using Hoobari The algorithm is used.

[0023] In one embodiment, determining the probability includes a fragment omics-based probability that the sequence read is of fetal origin.

[0024] In one embodiment, the method includes extracting fragmentomics features for each cfDNA read identified as overlapping with a potential genomic site that may have STR variants in the fetus.

[0025] In some embodiments, the fragmentomics features include one or more of the following: read quality mapping, read base quality, fragment length, short / long read ratio, terminal motif, cleavage pattern around methylation sites, preferred ends of reads, DNA / accessibility / nucleosome, distance to the nearest nucleosome, transcription factor binding sites, regional fetal fraction, regional sequence composition, read sequence composition, and the number of sequence errors in the reads.

[0026] In one embodiment, determining the probability includes multiplying the probability based on mutation detection by the probability based on fragment omics to obtain a calculated joint probability.

[0027] In some embodiments, the step of inferring maternal and optionally paternal STR alleles includes applying a bioinformatics tool selected from the group consisting of: STRetch, GangSTR, ExpansionHunter, HipSTR, Expansion Hunter De Novo (EHdn), STRling, RepeatSeq, LobSTR, adVNTR, and TredParse.

[0028] In some embodiments, the STR variant indicates a genetic disease selected from the group consisting of: Fragile X syndrome, Fragile X-related tremor / ataxia syndrome (FXTAS), X-linked intellectual disability, Huntington's disease, myotonic dystrophy, spinocerebellar ataxia (SCA), Friedreich ataxia, Machado-Joseph disease, spinocerebellar ataxia type 10 (SCA10), spinocerebellar ataxia type 12 (SCA12), spinocerebellar ataxia type 17 (SCA17), dentate nucleus-globus pallidus Lewy body atrophy (DRPLA), spinocerebellar ataxia type 8 (SCA8), spinocerebellar ataxia type 31 (SCA31), and Fragile XE syndrome.

[0029] In another aspect, the present invention provides a method for diagnosing genetic diseases in a fetus that are associated with or caused by short tandem repeat sequences (STRs), the method comprising: a. Receive sequencing data reads from (i) cell-free maternal plasma DNA (cfDNA) and (ii) maternal and optionally paternal genomic DNA (gDNA) from a pair of parents that nurtured the fetus; b. Identify potential genomic loci in the fetus that may have short tandem repeat (STR) variations; c. For each of the potential genomic loci, determine the probability that the fetus has the STR variant; d. Infer the STR alleles of the mother and the arbitrarily chosen father; e. Process cfDNA reads in regions of the gene that potentially contain STRs to identify read categories, wherein the read categories include spanning reads, partial reads (including indicative and non-indicative reads), and clean reads; f. Calculate the expected ratio between the read segment categories to obtain the calculated ratio; g. Determine the most likely inherited alleles of the fetus using the following method: i) Measure the observation ratio between the read segment categories; and ii) Use the calculated ratio and the observed ratio to distinguish normal alleles from alleles with premutation or full mutation, and to distinguish alleles with premutation from alleles with full mutation. or iii) Apply Bayesian algorithm; This allows for the diagnosis of the fetus.

[0030] In some embodiments, the genetic diseases associated with or caused by short tandem repeats (STRs) are selected from the group consisting of: Fragile X syndrome, Fragile X-related tremor / ataxia syndrome (FXTAS), X-linked intellectual disability, Huntington's disease, myotonic dystrophy, spinocerebellar ataxia (SCA), Friedreich ataxia, Machado-Joseph disease, spinocerebellar ataxia type 10 (SCA10), spinocerebellar ataxia type 12 (SCA12), spinocerebellar ataxia type 17 (SCA17), dentate nucleus-globus pallidus Lewy body atrophy (DRPLA), spinocerebellar ataxia type 8 (SCA8), spinocerebellar ataxia type 31 (SCA31), and Fragile XE syndrome.

[0031] In another aspect, the present invention provides a method for diagnosing fragile X syndrome in a fetus, the method comprising: a. Receive sequencing data reads of (i) cell-free DNA from maternal plasma (cfDNA) and (ii) maternal genomic DNA (gDNA); b. Inferring the maternal FMR1 gene alleles; c. Calculate the fetal fraction within the cfDNA; d. Process cfDNA reads in regions of the FMR1 gene to identify read categories, wherein the read categories include spanning reads, partial reads (including indicative and non-indicative reads), and clean reads; e. Calculate the expected ratio between the read segment categories to obtain the calculated ratio; f. Determine the most likely inherited alleles of the fetus by the following method: i. Measure the observation ratio between the read segment categories; and ii. Use the calculated ratio and the observed ratio to distinguish normal alleles from alleles with premutations or full mutations, and to distinguish alleles with premutations from alleles with full mutations; or iii. Apply Bayesian algorithm; in An imbalance toward crossing reading segments indicates a high probability that the fetus does not have Fragile X syndrome. An imbalance toward the indicative read indicates a high probability that the fetus has a premutation or full mutation in the FMR1 gene, and An imbalance toward the indicative read and a ratio between clean reads and indicative reads that is higher than the calculated ratio indicate a high probability that the fetus has a full mutation in the FMR1 gene that leads to Fragile X syndrome.

[0032] In one embodiment, the step of calculating the expected ratio between the read segment categories includes normalizing the read segment counts.

[0033] In some embodiments, the normalization is performed by one or more of the following: a. Intra-sample normalization b. Sample normalization c. Comparison of fetal enrichment and maternal enrichment reading segments. d. Haplotype comparison, and e. Correction using two-terminal NGS technology.

[0034] In one embodiment, the inter-sample normalization includes creating a normal sample set (PON).

[0035] In another aspect, the present invention provides a computer software product comprising a computer-readable medium therein storing program instructions that, when read by a data processor, configure the data processor to: (1) receive sequencing data reads of (i) maternal cell-free DNA (cfDNA) and (ii) maternal and optionally paternal genomic DNA (gDNA) from a pair of parents of a developing fetus, and (2) perform the method of the present invention.

[0036] In another aspect, the present invention provides a system for genotyping a fetus, the system comprising: an input utility for receiving sequencing data reads from a pair of parents of the fetus, namely (i) maternal cell-free DNA (cfDNA) and (ii) maternal and optionally paternal genomic DNA (gDNA); and a data processor configured to analyze the data to perform the method of the present invention. Attached Figure Description

[0037] To better understand the subject matter disclosed herein and to illustrate how it is implemented in practice, embodiments will now be described by way of non-limiting examples only, with reference to the accompanying drawings, in which: Figure 1 Here is a summary of a method for fetal genotyping according to various exemplary embodiments of the present invention.

[0038] Figure 2 Examples include STR-containing genes and their aligned read categories.

[0039] Figure 3 Here is an outline of a method applicable to fetal genotyping, particularly for identifying amplification mutations associated with Fragile X syndrome. Detailed Implementation

[0040] This invention provides methods and systems for predicting the risk of genetic amplification mutations, for distinguishing premutated fetuses from healthy fetuses with high probability, and for estimating the risk of genetic full mutations. The methods and systems of this invention are applicable to all types of amplification mutations. Non-limiting examples of the methods and systems of this invention relate to identifying amplification mutations associated with Fragile X syndrome.

[0041] According to the present invention, maternally derived trinucleotide amplification mutations in the fetus are detected via NIPT.

[0042] The method of this invention can be performed using long read and / or short read NGS techniques. Note that although the method according to this invention can use linked read sequencing or long cfDNA molecules, there are no requirements regarding linked read sequencing or long cfDNA molecules.

[0043] The method of this invention is direct, meaning it detects the amplification itself rather than the markers surrounding the amplification. This method may optionally be combined with phasing of the parental genome to improve prediction accuracy.

[0044] As used in this article, terminology "Pre-mutation" (PM) refers to a repeat range that undergoes amplification but does not cause the underlying disease. For example, in the FMR1 gene (the gene that causes Fragile X syndrome), 5-44 is the normal repeat range, 45-54 is the gray area (also known as the intermediate allele), and 55-200 is considered a premutation. Between 55-200 repeats, amplification occurs in the next generation, which itself does not cause Fragile X syndrome, although it can cause other phenotypes in females. Therefore, it is called a premutation. Those in their premutation (PM) state... FMR1 Genes have already been FMR1 A range of clinical and subclinical phenotypes, including some subclinical features associated with autism spectrum disorder (ASD), are associated with PM carriers. A mutation is considered complete when there are more than 200 repeats.

[0045] As used in this article, terminology "Full mutation" This refers to the extent to which the amplification of the repeating region exceeds the threshold defined by the disease. For example, in Fragile X syndrome, this corresponds to 200 repeats.

[0046] Short tandem repeat mutations are typically characterized by genotype-phenotype correlation. Therefore, within a certain range of repetitions, the more repetitions there are, the more severe the phenotype usually is, or the greater the likelihood of a complete mutation.

[0047] As used in this article, terminology "Early appearance of genetic traits" This refers to the phenomenon where, due to the increasing proliferation across generations, the disease appears at an earlier age in each generation.

[0048] Therefore, in one aspect, the present invention provides a method for genotyping a fetus, comprising: a. Receive sequencing data reads from (i) cell-free maternal plasma DNA (cfDNA) and (ii) maternal and optionally paternal genomic DNA (gDNA) from a pair of parents that nurtured the fetus; b. Identify potential genomic loci in the fetus that may have short tandem repeat (STR) variations; c. For each of the potential genomic loci, determine the probability that the fetus has the STR variant; d. Infer the STR alleles of the mother and the arbitrarily chosen father; e. Process cfDNA reads in regions of the gene that potentially contain STRs to identify read categories, wherein the read categories include spanning reads, partial reads, and clean reads; f. Calculate the expected ratio between the read segment categories to obtain the calculated ratio; g. Determine the most likely inherited alleles of the fetus using the following method: i. Measure the observation ratio between the read segment categories; and ii. Use the calculated ratio and the observed ratio to distinguish normal alleles from alleles with premutations or full mutations, and to distinguish alleles with premutations from alleles with full mutations; or iii. Apply Bayesian algorithm; This allows for genotyping of the fetus.

[0049] Figure 1 An exemplary summary of an embodiment of the method of the present invention for fetal genotyping is shown.

[0050] First, obtain parental (maternal and optionally paternal) and cfDNABAM files containing parental and fetal-maternal reads respectively, as well as FF (e.g., as shown in the image). Figure 1 (10% as shown) (steps (a)-(c) in the above method). Parental alleles are inferred from the parental BAM (step (d)), and the reads corresponding to the query gene are processed from the cfDNA BAM (step (e)). The expected ratio between the cross-reads originating from known normal parental alleles and the indicative reads originating from the premutation is calculated based on FF. The expected ratio is corrected for biases and errors caused by biological, technical, or any other systematic biases and errors (step (f)). Based on the observed ratio, it is determined whether the fetus has inherited the normal allele; the probability of a full mutation is predicted based on the ratio of clean reads to other reads (e.g., indicative reads) (steps (g)(i) and (ii)).

[0051] Also known as "Circulating cell-free DNA" Cell-free DNA (cfDNA) is found in bodily fluids such as blood plasma. in vivo Circulating fetal DNA (cfDNA) is a fragment of DNA that exists outside the cell. The length of cfDNA fragments is typically between 150 and 200 base pairs (bp), with an average of about 170 bp, which may be related to the length of the DNA extension surrounding the nucleosome. During pregnancy, cell-free fetal DNA can be found circulating in the maternal plasma. Therefore, cfDNA in maternal plasma is a mixture of maternal and fetal DNA; both the total amount of cfDNA and the fraction of fetal DNA within it increase throughout pregnancy.

[0052] The term cfDNA also refers to DNA from in vivo extracellular source acquisition and In vitro Fragments of DNA that have been isolated, separated, or otherwise processed. cfDNA can be obtained by extracting DNA from plasma after removing intact cells. Methods for extracting cfDNA are well known in the art, for example, as illustrated in the examples below.

[0053] the term "Genomic DNA" or "gDNA" In this article, it refers to the existence of in vivo DNA within a cell and containing the complete genome of the cell or organism. The term also refers to DNA derived from... in vivo Obtained in cells and In vitro DNA is separated, isolated, or otherwise processed. Typically, cells are separated to produce DNA before undergoing lysis. in vitro Cellular DNA. As used herein, the term gDNA does not include cfDNA.

[0054] Maternal genomic DNA (gDNA) data, maternal cell-free DNA (cfDNA) data, and optionally paternal gDNA data are obtained through sequencing methods, including but not limited to deep whole-genome sequencing (WGS), whole-exome sequencing (WES), next-generation sequencing (NGS), targeted sequencing, kit sequencing, gene sequencing, long-read genome sequencing, paired-end sequencing, single-end sequencing, and amplicon sequencing.

[0055] the term Next-generation sequencing In this paper, (NGS) refers to sequencing methods that allow for massively parallel sequencing of cloned amplified molecules and individual nucleic acid molecules. Non-limiting examples of NGS include sequencing-by-synthesis using reversible dye terminators and sequencing-by-ligation.

[0056] Deep sequencing refers to sequencing a region of the genome multiple times, sometimes hundreds or even thousands of times. Deep sequencing of the genome enables researchers to detect rare genetic variations.

[0057] As used in this article Deep whole-genome sequencing This refers to deep sequencing of the entire genome. In the context of this invention, deep whole-genome sequencing is performed on cell-free DNA extracted from maternal plasma during pregnancy. Maternal plasma samples can be obtained at any stage of pregnancy (preferably between the 7th and 38th weeks of gestation).

[0058] Sequencing is repeated multiple times, for example, but not limited to, between 10 times (10X) and 1000 times (1000X), such as 10 times (10X), 20 times (20X), 30 times (30X), 50 times (50X), 100 times (100X), 150 times (X150), 200 times (200X), 300 times (300X), 500 times (500X), or 1000 times (1000X).

[0059] Plasma cfDNA can be sequenced at different depths.

[0060] In one non-limiting example, the cfDNA in maternal plasma was sequenced 300 times (300X); in other embodiments, the cfDNA in maternal plasma was sequenced 50 times (50X), 100 times (100X), or 150 times (X150).

[0061] In addition, whole-genome sequencing is performed on both the maternal and optional paternal genomic DNA. This genomic DNA can be obtained from any cell type (e.g., from blood cells, such as white blood cells). In this embodiment, whole-genome sequencing of the paternal and maternal genomic DNA is performed to a target depth of approximately 20-fold to 40-fold (e.g., 30-fold).

[0062] Whole genome sequencing can be performed using any method known in the art, such as the HiSeq X Ten system (Illumina), HiSeq4000 (Illumina), nanopore WGS sequencing using the MinION device (Oxford Nanopore Technologies), and WGS from Ultima Genomics.

[0063] Sequencing generation "Read the paragraph" A read is a sequence of DNA fragments of varying lengths. Typically, though not always, a read represents a short sequence of consecutive bases in a sample. Reads can be symbolically represented by base pair sequences (in ATCG). They can be stored in storage devices and processed as needed. Reads can be obtained directly from the sequencing device or indirectly from stored sequence information.

[0064] Sequencing inputs can be long reads (e.g., from about 1 kBP to about 100 kBP or more) or short reads (e.g., from about 50 to 400 base pairs), or a combination of long and short reads.

[0065] After sequencing, the reads are compared with the human reference genome based on sequence similarity.

[0066] The identification of maternal and paternal variations (i.e., variant sites or mutations) can be performed using variation detection methods, which are typically based on the comparison of DNA sequencing data and the application of commercially available variation detectors.

[0067] As used in this article, terminology "aligned" , "alignment" or "aligning" Alignment refers to the process of comparing a read segment with a reference sequence and thus determining whether the read segment is contained within the reference sequence. If the reference sequence contains the read segment, the read segment can be mapped to a specific position within the reference sequence. In some cases, alignment simply indicates whether the read segment exists in the reference sequence.

[0068] Sequence alignment techniques that can be used according to some embodiments of the present invention include, but are not limited to: BurrowsWheeler Aligner (BWA), ABA, ALE, AMAP, anon, BAli-Phy, Base-By-Base, BHAOS / DIALIGN, Bowtie, Bowtie 2, ClustalW, CodonCode Aligner, Comass, DECIPHER, DIALIGN-TX, DIALIGN-T, DNA Alignment, and DNA Baser Sequence. Assembler, EDNA, FSA, Geneious, Kalign, MAFFT, MARNA, MAVID, MSA, MSAProbs, MULTALIN, Multi-LAGEN, MUSCLE, Opal, Pecan, Phylo, Praline, PicXAA, POA, Probalign, P robCons, PROMALS3D, PRRN / PRRP, PSAlign, RevTrans, SAGA, SAM, Se-AI, STAR, STAR-Fusion, StatAlign, Stemloc, T-Coffee, UGENE, VectorFriends, NovoAlign and GLProbs.

[0069] Exemplary variant detectors applicable to this embodiment include, but are not limited to, the Genome Analysis Toolkit (GATK) and Freebayes. For example, Freebayes may include alignments of text sequences based on reads aligned to a specific target, rather than their exact alignments. GATK may include: (i) preprocessing; (ii) variant discovery; and (iii) detection set optimization. Preprocessing may include starting with raw sequence data (e.g., FASTQ or uBAM format) and producing an analysis-ready BAM file; processing may include alignment with a reference genome and data cleaning operations to correct for technical biases and make the data suitable for analysis; variant discovery may include starting with an analysis-ready BAM file and producing a detection set in VCF format; processing may involve identifying sites in one or more individuals that show possible genomic variants and applying filtering methods suitable for the experimental design; detection set optimization may include starting and ending with a VCF detection set; processing may involve using metadata to assess and improve genotyping accuracy, adding additional information, and evaluating the overall quality of the detection set.

[0070] Variant detectors are also considered, such as, but not limited to, Platypus, VarScan, Bowtie analysis, MuTect, Google DeepVariant, and / or SAMtools. For example, Bowtie analysis may include implementing the Burrows-Wheeler transform for alignment. MuTect may include: (i) preprocessing; (ii) statistical analysis; and (iii) postprocessing. Preprocessing may include initial alignment of sequencing reads; statistical analysis may include using two Bayesian classifiers, one classifier that detects whether an SNP is a non-reference site at a given site, and another classifier that ensures normal sites do not carry SNPs for those sites found to be non-reference sites; postprocessing may include removing artifacts from sequencing, short read alignment, and hybridization capture. SAMtools may include storing, manipulating, and aligning sequencing reads stored as SAM files.

[0071] In this embodiment, the step of determining the probability that a sequence read in maternal plasma cfDNA is of fetal origin is performed, for example, by an algorithm (e.g., the algorithm described in Rabinowitz et al., 2019 and US2021 / 0340601). Hoobari This process is used to calculate the fetal fraction (FF), which is the percentage of fetal-derived cfDNA in maternal plasma cfDNA, and to calculate the fragment length distribution. This step may also include extracting features at various fragment levels.

[0072] In various exemplary embodiments of the present invention, the method includes determining the probability that each DNA fragment (or read) is of fetal origin. Using these probabilities at the read level, genotype prediction is performed, i.e., predicting whether the fetus carries a normal allele, a premutation, or a full mutation.

[0073] In an embodiment, determining the probability that a DNA fragment (or read) is of fetal origin includes: constructing a fetal size distribution and a maternal size distribution, binning the fetal size distribution and calculating the fetal score for each fragment size bin, and for at least one size and at least one fragment at the at least one locus, calculating the probability that the fragment is fetal based on the fetal score of the corresponding fragment size bin to which the fragment belongs.

[0074] As used in this article "Fetal score" or "FF" This refers to a portion of the total fetal cfDNA within the mother's bloodstream. The fraction of fetal cfDNA in the mother's bloodstream (fetal fraction) varies throughout pregnancy and between individuals; therefore, it is not considered a constant but a variable. Low levels of fetal cfDNA are referred to as a low fetal fraction.

[0075] In an embodiment, determining the probability includes applying a Bayesian procedure. Optionally, the Bayesian procedure includes prior probabilities calculated using sequencing data from at least one of the parents.

[0076] In an embodiment, this procedure further includes a recalibrated output of the Bayesian procedure using machine learning.

[0077] In a specific embodiment, determining the probability that each DNA fragment (or read) is of fetal origin is done using variant detection (e.g., using methods described in Rabinowitz et al., 2019 and US2021 / 0340601). Hoobari It is performed using an algorithm.

[0078] Variance detection prediction can optionally be combined with the analysis of fragmentomics characteristics of DNA reads.

[0079] the term "Fragment omics characteristics" This refers to the molecular characteristics of a DNA read, as well as its genomic, epigenetic, and alignment features. Fragmentomics features include, but are not limited to: read quality mapping, read base quality, fragment length, short / long read ratio, DNA fragment terminal motifs, cleavage patterns around methylation sites, preferred ends of reads, DNA accessibility / nucleosomes, distance to the nearest nucleosome, transcription factor binding sites, regional fetal fraction, regional sequence composition, read sequence composition, and the number of sequence errors in the read.

[0080] In some embodiments, the mutation detection probability is multiplied by the fragment omics-based probability to calculate the joint probability, for which maximum likelihood is applied to predict the genotype at each locus.

[0081] The step of inferring maternal and arbitrarily paternal STR alleles involves calculating the number of repetitions in each maternal and arbitrarily paternal haplotype.

[0082] Inference of maternal and optionally paternal STR alleles can be made using any suitable bioinformatics tool, including but not limited to STRetch (Dashnow et al.), GangSTR, ExpansionHunter (Dolzhenko et al.), HipSTR, Expansion Hunter De Novo (EHdn), STRling, RepeatSeq, LobSTR, adVNTR, or TredParse.

[0083] The cfDNA reads in the region of the queried gene (i.e., the gene that potentially contains the STR) are processed and classified as follows: (i) “Spanning reads” – containing the full length of normal alleles.

[0084] (ii) “Partial reads” – containing a portion of a repetitive region and a portion of an adjacent genomic region. This group is further divided into: a. Upstream and downstream – This classification helps reduce noise and bias.

[0085] b. Indicative and Non-indicative—Indicative reads contain more repeats than the number of repeats defined for a normal allele, therefore they must originate from a premutated allele. Non-indicative reads contain fewer repeats than the threshold defined for a normal allele, and therefore can originate from either a normal or premutated allele.

[0086] As indicated above, there are three allele types: normal (i.e., with 44 or fewer repeats), premutation (i.e., with 45-200 repeats), and full mutation (i.e., with >200 repeats).

[0087] Therefore, indicative reads are partial reads with 45 or more repeats, and thus they must originate from a premutation (and are therefore indicative). Partial reads with 44 or fewer repeats can originate from normal alleles or from premutated alleles, and are therefore called non-indicative.

[0088] (iii) "Clean readings" - containing only repetitions.

[0089] These reading categories are in Figure 2The diagram illustrates the general structure of a relevant read at an exemplary region containing an STR gene. In this example, the normal allele consists of 40 repeats, and the premutated allele consists of 150 repeats. Cross-reads contain repeating regions and non-repeating flanking regions. Because premutated alleles are typically large, cross-reads usually correspond only to the normal allele unless long-read sequencing techniques are used. Clean reads correspond to either the normal or premutated allele. Indicative partial reads can only originate from the premutated allele, while non-indicative partial reads can originate from either allele.

[0090] The expected ratio between different reading segments (also referred to in this article) "Calculate the ratio" The ratio depends on the fetal fraction (FF), the inherited alleles, and the extension length (if a mutation has occurred). Additionally, it depends on different biases and error patterns because reads are not uniformly distributed across the queried region. Therefore, in very simple scenarios, the theoretically calculated / expected ratio might be 1, but this value can vary slightly depending on read length and STR length. Different physical and chemical properties between genomic regions can also cause deviations in this ratio.

[0091] Therefore, a normalization step is adopted. As will be explained below, it is possible to... Compared to Normalize other samples, other regions within a sample, or subsets of reads within the same region (two haplotypes or fetal and maternal enrichment).

[0092] Therefore, in order to estimate the expected / calculated ratio between read types, read counts are normalized relative to different references using various methods. These include, but are not limited to, one or more of the following: - Intrasample normalization - Normalization is performed on similar regions in the genome using GC content, comparability, and other unique measurements that are particularly relevant to the STRs being tested, such as the count of identical triplets in other genomic regions, to create a matrix of similar regions.

[0093] - Inter-sample normalization - Create normal sample sets (PON) or anomalous sample sets to understand typical ratios and their correlation with extension length.

[0094] - Comparison of fetal enriched and maternally enriched reads - cfDNA was separated into fetal enriched and maternally enriched reads (using methods described, for example, in Rabinowitz et al., 2019, US2021 / 0340601, and IL307784) and the same phenomenon was tested in both groups, or whether the phenomenon was more abundant in the fetal group.

[0095] - Haplotype comparison - Phase (i.e., separate haplotypes) of cfDNA sequencing data using parental (i.e., maternal and optionally paternal) haplotype information and verify whether the ratios between the two haplotypes are similar or different.

[0096] - Correction of paired-end NGS technology - Perform correction of paired-end NGS technology, that is, model the paired reads to fine-tune the expected ratio.

[0097] Determining the most likely inherited allele in a fetus can be done by measuring the observed ratios between different categories and comparing the observed ratios with the expected ratios. Alternatively or additionally, determining the most likely inherited allele in a fetus can be done by applying a Bayesian algorithm.

[0098] Measuring the observed ratio between different read classes can be used to distinguish normal alleles from premutations or full mutations, for example: if the mother carries a premutation and the father has two normal copies, the ratio between upstream (or downstream) indicative partial reads and spanned reads is expected to be 1:1 when maternally derived cfDNA is involved. An imbalance occurs if the fetus inherits a normal allele. On the other hand, if only the father is a carrier, the presence of indicative partial reads in maternal plasma will indicate inherited mutations.

[0099] Measuring the observed ratio between different read categories can also be used to distinguish between premutations and complete mutations, for example, by measuring the number of clean reads. A significant increase in this number compared to the expected ratio indicates a complete mutation. This calculation is also related to fetal scores.

[0100] Alternatively, Bayesian algorithms (such as those described in Rabinowitz et al., 2019 and US2021 / 0340601) can be applied to determine the most likely alleles inherited by the fetus, for example, as follows: a. Determine the possible fetal genotypes based on the parental genotypes. For example, if the mother has 20 and 30 repeated alleles, and the father has 25 and 31 repeated alleles, the fetus could be (20,25), (20,31), (30,25), or (30,31), and the prior probability of each fetal genotype is exactly 25%.

[0101] b. Use a suitable, readily available commercial software tool to calculate the probability that each read supports each allele, such as ExpansionHunter as described by Dolzhenko et al., HipSTR (haplotype inference and phasing of short tandem repeats), or tools as described by Liautard-Haag et al., and calculate the probability that each read originates from the fetus or mother as described in Rabinowitz et al., 2019 or US2021 / 0340601.

[0102] c. Calculate the probability that each read supports each genotype based on the probability that each allele constituting the genotype is supported by the read, and the probability that the read is of fetal or maternal origin.

[0103] d. The joint probability (proportional to the most likely genotype) is obtained by calculating the product of the probabilities of all reads for each genotype and multiplying this value by the prior probability of each genotype.

[0104] Therefore, the genotype with the highest probability of combination was detected.

[0105] In this embodiment, parental (maternal and optionally paternal) haplotype information is incorporated into the method. Parental sequencing information is phased based on short-read or long-read data (see, for example, PCT / IL2023 / 051184). Thus, indirect genotyping is added to the method using haplotypes around two alleles. This haplotype information is integrated with a Bayesian algorithm and can improve the accuracy of genotyping.

[0106] The method of this invention is applicable to the identification of human genetic diseases associated with or caused by short tandem repeats (STRs). Non-limiting examples of such genetic diseases include Fragile X syndrome (FMR1 gene), Fragile X-related tremor / ataxia syndrome (FXTAS) (FMR1 gene), X-linked intellectual disability (multiple genes, e.g., ARX, FMR1), Huntington's disease (HTT gene), myotonic dystrophy (DMPK gene), spinocerebellar ataxia (SCA) (multiple genes, e.g., SCA1, SCA2, SCA3), Friedreich ataxia (FXN gene), and Machado-Joseph disease (ATXN3 gene). Spinocerebellar ataxia type 10 (SCA10) (ATTCT repeat amplification), spinocerebellar ataxia type 12 (SCA12) (CAG repeat amplification), spinocerebellar ataxia type 17 (SCA17) (TBP gene), dentate nucleus-rubella-globus pallidus Lewy body atrophy (DRPLA) (ATN1 gene), spinocerebellar ataxia type 8 (SCA8) (CTG repeat amplification), spinocerebellar ataxia type 31 (SCA31) (BEAN1 gene), and fragile XE syndrome (AFF2 gene).

[0107] In one non-limiting embodiment, the present invention provides a method for genotyping a fetus, the method comprising identifying whether the fetus contains STR duplications in the FMR1 gene that may lead to Fragile X syndrome. That is, the present invention provides a non-invasive method for diagnosing Fragile X syndrome in a fetus.

[0108] This method is applicable to diagnosing Fragile X in both male and female fetuses.

[0109] Fragile X syndrome (FXS) is an X-linked genetic disorder and the most common known cause of inherited intellectual disability, as well as the main single mutation leading to autism spectrum disorder. Figure 3 An exemplary outline of an embodiment of the present invention for detecting STR duplications associated with Fragile X syndrome is shown.

[0110] Since this disease is associated only with the X chromosome, the analysis according to this embodiment of the invention focuses only on male fetuses. These fetuses inherit one X-linked allele. Therefore, the paternal allele is not described. The fetus inherits only one maternal allele, meaning that all FF (5% in this example) correspond to this allele. Therefore, an imbalance toward cross-reads can exclude affected fetuses. An imbalance toward indicative reads means that affected fetuses cannot be excluded, and the ratio between indicative and clean reads helps estimate the risk of a full mutation. That is, a ratio between clean and indicative reads higher than the calculated ratio indicates the presence of a full mutation.

Claims

1. A method for genotyping a fetus, the method comprising: a. Receive sequencing data reads from (i) cell-free maternal plasma DNA (cfDNA) and (ii) maternal and optionally paternal genomic DNA (gDNA) from a pair of parents that nurtured the fetus; b. Identify potential genomic loci in the fetus that may have short tandem repeat (STR) variations; c. For each of the potential genomic loci, determine the probability that the fetus has the STR variant; d. Infer the STR alleles of the mother and the arbitrarily chosen father; e. Process cfDNA reads in regions of the gene that potentially contain STRs to identify read categories, wherein the read categories include spanning reads, partial reads (including indicative and non-indicative reads), and clean reads; f. Calculate the expected ratio between the read segment categories to obtain the calculated ratio; g. Determine the most likely inherited alleles of the fetus using the following method: i. Measure the observation ratio between the read segment categories; and ii. Use the calculated ratio and the observed ratio to distinguish normal alleles from alleles with premutations or full mutations, and to distinguish alleles with premutations from alleles with full mutations; or iii. Apply Bayesian algorithm; This allows for genotyping of the fetus.

2. The method according to claim 1, wherein the sequencing data reads are short reads.

3. The method according to claim 1, wherein the sequencing data read is a long read.

4. The method according to claim 1, wherein the sequencing data reads comprise short reads and long reads.

5. The method according to any one of claims 1 to 4, wherein one or both of the gDNA sequencing data and cfDNA sequencing data are obtained by a method selected from the group consisting of: whole genome sequencing (WGS), whole exome sequencing (WES), next-generation sequencing (NGS), targeted sequencing, kit sequencing, gene sequencing, long read genome sequencing, paired-end sequencing, single-end sequencing, and amplicon sequencing.

6. The method of claim 5, wherein the WGS or WES data is obtained through deep sequencing.

7. The method according to any one of the preceding claims, wherein the probability is determined based on at least one sequence alignment map (SAM) parameter.

8. The method according to any one of the preceding claims, wherein the probability is determined based at least on the observed template length.

9. The method according to any one of the preceding claims, wherein determining the probability includes calculating a total fetal score.

10. The method of claim 9, further comprising constructing a fetal size distribution and a maternal size distribution, wherein determining the probability in step c of claim 1 comprises: The fetal size distribution is binned and the fetal score of each segment size bin is calculated. For at least one size and at least one segment at the at least one location, the probability that the segment is a fetus is calculated based on the fetal score of the corresponding segment size bin to which the segment belongs.

11. The method according to any one of the preceding claims, wherein determining the probability comprises applying a Bayesian procedure.

12. The method of claim 11, wherein the Bayesian procedure includes prior probabilities calculated using sequencing data of at least one of the parents.

13. The method of claim 11 or 12, further comprising recalibrating the output of the Bayesian procedure using machine learning.

14. The method according to any one of the preceding claims, wherein the determination of the probability is performed using fetal variation detection.

15. The method according to any one of the preceding claims, wherein determining the probability is done using... Hoobari The algorithm is used.

16. The method according to any one of the preceding claims, wherein determining the probability includes a fragment omics-based probability that the sequence read is of fetal origin.

17. The method of claim 16, wherein the method comprises extracting fragmentomics features for each cfDNA read identified as overlapping with a potential genomic site that may have STR variants in the fetus.

18. The method of claim 17, wherein the fragmentomics features include one or more of the following: read quality mapping, read base quality, fragment length, short / long read ratio, terminal motif, cleavage pattern around methylation sites, preferred ends of reads, DNA / accessibility / nucleosome, distance to the nearest nucleosome, transcription factor binding sites, regional fetal fraction, regional sequence composition, read sequence composition, and the number of sequence errors in the reads.

19. The method of any one of claims 16 to 18, wherein determining the probability comprises multiplying the probability based on mutation detection by the probability based on fragment omics to obtain a calculated joint probability.

20. The method according to any one of the preceding claims, wherein the step of inferring maternal and optionally paternal STR alleles comprises applying a bioinformatics tool selected from the group consisting of: STRetch, GangSTR, ExpansionHunter, HipSTR, Expansion Hunter De Novo (EHdn), STRling, RepeatSeq, LobSTR, adVNTR, and TredParse.

21. The method according to any one of the preceding claims, wherein the STR variant indicates a genetic disease selected from the group consisting of: Fragile X syndrome, Fragile X-related tremor / ataxia syndrome (FXTAS), X-linked intellectual disability, Huntington's disease, myotonic dystrophy, spinocerebellar ataxia (SCA), Friedreich ataxia, Machado-Joseph disease, spinocerebellar ataxia type 10 (SCA10), spinocerebellar ataxia type 12 (SCA12), spinocerebellar ataxia type 17 (SCA17), dentate nucleus-globus pallidus Lewy body atrophy (DRPLA), spinocerebellar ataxia type 8 (SCA8), spinocerebellar ataxia type 31 (SCA31), and Fragile XE syndrome.

22. A method for diagnosing genetic diseases in a fetus associated with or caused by short tandem repeats (STRs), the method comprising: a. Receive sequencing data reads from (i) cell-free maternal plasma DNA (cfDNA) and (ii) maternal and optionally paternal genomic DNA (gDNA) from a pair of parents that nurtured the fetus; b. Identify potential genomic loci in the fetus that may have short tandem repeat (STR) variations; c. For each of the potential genomic loci, determine the probability that the fetus has the STR variant; d. Infer the STR alleles of the mother and the arbitrarily chosen father; e. Process cfDNA reads in regions of the gene that potentially contain STRs to identify read categories, wherein the read categories include spanning reads, partial reads (including indicative and non-indicative reads), and clean reads; f. Calculate the expected ratio between the read segment categories to obtain the calculated ratio; g. Determine the most likely inherited alleles of the fetus using the following method: i) Measure the observation ratio between the read segment categories; and ii) Use the calculated ratio and the observed ratio to distinguish normal alleles from alleles with premutation or full mutation, and to distinguish alleles with premutation from alleles with full mutation. or iii) Apply Bayesian algorithm; This allows for the diagnosis of the fetus.

23. The method of claim 22, wherein the genetic disease associated with or caused by short tandem repeats (STRs) is selected from the group consisting of: Fragile X syndrome, Fragile X-related tremor / ataxia syndrome (FXTAS), X-linked intellectual disability, Huntington's disease, myotonic dystrophy, spinocerebellar ataxia (SCA), Friedreich ataxia, Machado-Joseph disease, spinocerebellar ataxia type 10 (SCA10), spinocerebellar ataxia type 12 (SCA12), spinocerebellar ataxia type 17 (SCA17), dentate nucleus-globus pallidus Lewy body atrophy (DRPLA), spinocerebellar ataxia type 8 (SCA8), spinocerebellar ataxia type 31 (SCA31), and Fragile XE syndrome.

24. A method for diagnosing fragile X syndrome in a fetus, the method comprising: a. Receive sequencing data reads of (i) cell-free DNA from maternal plasma (cfDNA) and (ii) maternal genomic DNA (gDNA); b. Inferring the maternal FMR1 gene alleles; c. Calculate the fetal fraction within the cfDNA; d. Process cfDNA reads in regions of the FMR1 gene to identify read categories, wherein the read categories include spanning reads, partial reads (including indicative and non-indicative reads), and clean reads; e. Calculate the expected ratio between the read segment categories to obtain the calculated ratio; f. Determine the most likely inherited alleles of the fetus by the following method: i. Measure the observation ratio between the read segment categories; and ii. Use the calculated ratio and the observed ratio to distinguish normal alleles from alleles with premutations or full mutations, and to distinguish alleles with premutations from alleles with full mutations; or iii. Apply Bayesian algorithm; in An imbalance toward crossing reading segments indicates a high probability that the fetus does not have Fragile X syndrome. An imbalance toward the indicative read indicates a high probability that the fetus has a premutation or full mutation in the FMR1 gene, and An imbalance toward the indicative read and a ratio between clean reads and indicative reads that is higher than the calculated ratio indicate a high probability that the fetus has a full mutation in the FMR1 gene that leads to Fragile X syndrome.

25. The method according to any one of the preceding claims, wherein the step of calculating the expected ratio between the read segment categories includes normalizing the read segment counts.

26. The method of claim 25, wherein the normalization is performed by one or more of the following: a. Intra-sample normalization b. Sample normalization c. Comparison of fetal enrichment and maternal enrichment reading segments. d. Haplotype comparison, and e. Correction using two-terminal NGS technology.

27. The method of claim 26, wherein the inter-sample normalization includes creating a normal sample set (PON).

28. A computer software product comprising a computer-readable medium storing program instructions which, when read by a data processor, configure the data processor to: (1) receive sequencing data reads of (i) maternal cell-free DNA (cfDNA) and (ii) maternal and optionally paternal genomic DNA (gDNA) from a pair of parents of a developing fetus, and (2) perform the method according to any one of claims 1 to 27.

29. A system for genotyping a fetus, the system comprising: An input utility for receiving sequencing data reads from a pair of parents of a developing fetus, namely (i) maternal cell-free DNA (cfDNA) and (ii) maternal and optionally paternal genomic DNA (gDNA); and a data processor configured to analyze said data to perform the method according to any one of claims 1 to 27.

Images