A method for assessing fetal DNA concentration (cffDNA) in maternal plasma through blocks of linkage disequilibrium at polymorphic loci

Abstract

A method for evaluating fetal DNA concentration (cffDNA) in maternal blood during pregnancy through polymorphic locus linkage block, comprising: obtaining low-depth WGS sequencing data of the sample; obtaining a mapping relationship between the locus and the linkage polymorphism; and calculating the concentration according to the low-depth WGS sequencing data and the mapping relationship between the locus and the linkage polymorphism and taking a plurality of loci constituting the linkage polymorphism as a unit. Wherein, the mapping relationship between the locus and the linkage polymorphism can be obtained by processing public genome WGS data, processing public linkage polymorphism block data or processing a large amount of high-depth WGS measured data. The present application is based on low-depth WGS data, and the heterozygosity of the linkage SNP block can be used to calculate the cffDNA concentration in the maternal blood during pregnancy, which is not affected by the fetal gender in the actual application of NIPT, does not require the father's sample and does not require additional experiments, has high stability, high performance and low cost.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of bio-information analysis, and particularly relates to a method for evaluating fetal DNA concentration (cffDNA) in maternal blood during pregnancy through linkage of polymorphic loci, which can be used for evaluating fetal DNA concentration in maternal blood during pregnancy based on low-depth WGS. BACKGROUND

[0002] The proportion of fetal cell free DNA (cffDNA) in maternal peripheral blood cell free DNA (cfDNA) during pregnancy, i.e. fetal concentration, is called cffDNA concentration (fetal fraction, FF). The cffDNA concentration is crucial for the accuracy of non-invasive prenatal testing (NIPT) results. At present, NIPT requires that the cffDNA concentration be at least 4%, and if the cffDNA concentration in maternal blood is <4%, the abnormal fetal DNA may be affected by a relatively large amount of maternal normal euploid DNA, so that the accurate results cannot be detected, resulting in false negative results.

[0003] Currently, there are several methods to detect the concentration of cffDNA in maternal plasma during the application of NIPT. Quantifying the concentration of cffDNA in maternal plasma from male fetuses based on Y-chromosome reads or genetic markers is the most accurate and direct method, but it is also the most flawed, only suitable for pregnant mothers carrying male fetuses. Some gender-independent methods rely on the different patterns between maternal and fetal cfDNA, such as differences in fragment length distribution, which are generally accurate; methods based on DNA methylation differences, which are expensive and lack stability; methods based on high-depth sequencing of maternal plasma DNA, which require high sequencing depth and are therefore costly; methods based on SNP sites of parental genotypes, which are direct and accurate, but require samples from the father of the fetus, which is not feasible in practical applications due to the fact that NIPT only collects samples from pregnant mothers; the SeqFF model developed by Sung K.Kim's team is a more accurate and efficient method, but this high-dimensional model requires large-scale sample data during construction, and its performance significantly decreases when FF is less than 5%. Existing technologies can detect fetal concentration using common NGS methods, such as the method for predicting whether a fetus carries a genetic defect disclosed in patent application No. CN202211280321.2 and the non-diagnostic method for non-invasive prenatal detection of deafness gene mutations disclosed in patent application No. CN202011338696.0. However, common NGS methods also have certain drawbacks: cfDNA in the peripheral blood of pregnant mothers contains a large amount of maternal DNA, with cffDNA accounting for only a small percentage, typically between 5% and 15%. Genetic mutations, copy number variations, and other conditions in the mother herself can interfere with the detection and analysis of cffDNA, affecting the accuracy of fetal DNA proportion calculations. Placental tissue may exhibit cytogenetic mosaicism, meaning that the genomic composition of placental cells does not completely match that of other fetal tissues. This can result in inaccurate results when detecting cffDNA from the placenta, which can affect the accuracy of cffDNA proportion detection. For the detection of low-frequency genetic mutations or other genetic variations in cffDNA, NGS methods have a certain false negative rate. When the proportion of cffDNA is low, these low-frequency variations may be overwhelmed by the large amount of normal DNA signal, making them difficult to accurately detect, which can affect the assessment of the true proportion of cffDNA. Therefore, the accuracy of NGS methods is not sufficient. In addition, regarding the detection range, cffDNA in blood typically exists in the form of short fragments, generally around 150-200 bp. NGS methods may have reduced detection efficiency for extremely short or extremely long DNA fragments, which can result in the ineffective detection of some cffDNA fragments, affecting the overall judgment of cffDNA proportion.There are some high GC content regions, repetitive sequence regions and the like in the genome. When the NGS technology is used to sequence these regions, problems such as uneven coverage and insufficient sequencing depth are prone to occur, and the cffDNA information in these regions may be missed, and the proportion of cffDNA in the whole genome cannot be accurately calculated. Further, the NGS detection process includes multiple links such as DNA extraction, library construction and sequencing, and strict quality control and standardized operation are required for each link. Various errors may be introduced in the operation process, such as loss in the DNA extraction process, preference in the library construction process and the like, which affect the detection of the proportion of cffDNA. Moreover, the data generated by NGS is huge, and multiple steps such as alignment, splicing and variation detection are required in the analysis process. The selection of algorithm and the setting of parameters may affect the analysis result, which increases the difficulty of accurately calculating the proportion of cffDNA, and the detection cost is high, the data typing credibility is low, and the maternal fetal signal cannot be directly separated.

[0004] For the sequencing method used in the evaluation of the concentration of fetal DNA in maternal blood during pregnancy, the existing ngs data used for prenatal non-invasive parentage identification are basically high-depth sequencing, and specific region information is captured by hybridization probes, and the sequencing depth is usually more than 200X. WGS sequencing does not have a specific region, and the range covers the whole genome. If high-depth sequencing is obtained, the sequencing cost is very high. For example, 30X wgs human genome sequencing requires about 90G of base sequencing, and the current sequencing cost is about 2000 yuan. Therefore, wgs sequencing has advantages over hybridization, reduces the operation process and saves time. The data typing credibility of low-depth wgs is low, and the fetal signal in the maternal cannot be directly separated. Therefore, how to use low-depth wgs data to evaluate the concentration of fetal DNA in maternal blood during pregnancy is a technical problem that needs to be solved in the field.

[0005] Based on the existing technology, there is now a need for a cffDNA detection method with lower detection cost, which can cope with low-quality samples and low-concentration sequencing. SUMMARY

[0006] Based on the above problems, the present application provides a method for evaluating the concentration of fetal DNA in maternal blood during pregnancy by using linkage block of polymorphic sites, which comprises the following steps:

[0007] S1: establishing a linkage relationship database of polymorphic sites;

[0008] S2: performing low-depth WGS sequencing on the sample to obtain sequencing data;

[0009] S3: screening the typing data of the corresponding sites constituting the linkage polymorphic site block on the reference genome in the sequencing data according to the linkage relationship of the polymorphic sites.

[0010] S4: performing gene detection according to the typing data and in units of linkage polymorphic site region blocks;

[0011] The linkage polymorphic site region block is a linkage polymorphic site region block composed of multiple linkage sites with high linkage degree, and the reference genome is public genome WGS data corresponding to a population or a population.

[0012] In the step S1, the linkage relationship of the polymorphic sites is obtained by at least one of the following methods: (1) obtained by processing public genome WGS data;

[0013] (2) obtained by processing a large amount of high-depth WGS measured data;

[0014] (3) obtained from haplotype cells, such as germ cell sequencing data;

[0015] (4) obtained by processing long sequencing mode, such as sanger, third-generation sequencing data;

[0016] (5) obtained by querying public databases or literature.

[0017] The processing of the public genome WGS data or the processing of the large amount of high-depth WGS measured data is specifically obtained by screening based on a reference genome according to a predetermined rule.

[0018] The terms used in the present application are defined as follows

[0019] The term "genome WGS data" refers to genome data obtained by whole genome sequencing technology, which contains sequence information of all bases in the genome.

[0020] The term "low-depth WGS (Whole Genome Sequencing)" refers to whole genome sequencing of a sample, but the sequencing depth is low, usually covering part of the region of the genome, which is used to obtain basic information of the genome.

[0021] The term "fetal cell-free DNA (Cell-Free Fetal DNA, cffDNA)": specifically refers to the free DNA fragments in the maternal blood during pregnancy, which are derived from the fetus, and are used for non-invasive prenatal testing (NIPT).

[0022] The term "Linkage Polymorphism" refers to a combination of polymorphisms of multiple loci in the genome that tend to be transmitted together during inheritance. In the present application, the calculation of fetal DNA concentration is based on the multiple loci that form a linkage polymorphism.

[0023] The term "Sequencing Depth" refers to the average number of times each base is sequenced during sequencing. Low-depth sequencing refers to a lower number of times each base is sequenced.

[0024] The term "Genome Reference Consortium Human Build 38 (GRCh38)" refers to a standard human genome reference sequence version released by the Human Genome Project. It is one of the most widely used reference genomes and is maintained and updated by the Genome Reference Consortium (GRC).

[0025] The term "SNP Region Block" refers to a unit collection of multiple linkage SNPs, including a core SNP site and one or more SNP sites linked to it.

[0026] The term "Linkage Region Block" refers to a unit collection of multiple polymorphic genetic marker sites with linkage relationships, including a core polymorphic site and one or more polymorphic sites linked to it.

[0027] In one specific embodiment of the present application, a method for evaluating fetal DNA concentration in maternal blood during pregnancy based on low-depth WGS is provided. However, the present application only supports the above-mentioned genetic detection use with this specific use, and is not intended to be limiting. In fact, those skilled in the art can reasonably infer other genetic detection uses of the same type.

[0028] The present method is exemplarily applied to non-invasive prenatal testing (NIPT) and is suitable for prenatal screening and inferring the chromosomal or genetic status of the fetus, rather than general disease diagnosis. Further, based on the detection results, the chromosomal abnormalities (such as Down syndrome, Edward syndrome, etc.) or other genetic diseases of the fetus can be detected, but other biomarkers or pathogen detection need to be combined. Therefore, the method of the present application is not a direct diagnostic method.

[0029] For individual precision medicine, disease diagnosis and other applications with high accuracy requirements, several hundred X or even higher sequencing depth is usually required, and low sequencing depth will greatly reduce the accuracy and reliability of the sequencing results. There may be more false negative results. When the sequencing depth is reduced to 0.25X, only a small part of the bases in the genome can be detected, and most of the regions may not be covered, and there will be a large number of sequencing blank areas. It is difficult to accurately detect the variation in the genome, because at low depth, even if there is a variation in a base, it may be detected only once or even less, and it is difficult to distinguish between real variation and sequencing error. Generally, only in the context of large-scale population genetics research, a sequencing depth as low as 0.25X can be used for preliminary screening to quickly obtain the approximate genetic information of the population at a lower cost, and to understand some common gene frequency distribution and other situations.

[0030] For the technical problem that the typing reliability of low-depth wgs data is low and cannot directly indicate the separation of the fetal signal in the maternal, the present application provides a solution to mine high-depth genetic variation types in wgs sequencing data, judge the genotype of the mother, and separate the fetal signal for concentration evaluation and parentage identification.

[0031] The present application explores the relationship between genetic variations, and the following uses SNP as an example.

[0032] For example, two sites in the CHB, CHS population of the 1000 Genomes Project,

[0033] chr4:213231:G>A rs5003260

[0034] chr4:221985:G>A rs6830496

[0035] The assembled genome has the following rules:

[0036] a. When the base at chr4:213231 is detected as G, the base at chr4:221985 must be detected as G

[0037] b. When the base at chr4:213231 is detected as A, the base at chr4:221985 must be detected as A.

[0038] Or under the limitation of not being able to assemble, the typing results also have the following rules:

[0039] a. When the genotype at chr4:213231 is detected as GG, the genotype at chr4:221985 must be detected as GG

[0040] b. If genotype AA is detected at chr4:213231, AA must be detected at chr4:221985

[0041] c. If genotype AG is detected at chr4:213231, AG must be detected at chr4:221985

[0042] Obviously, extending this situation, the same rule applies to third-generation sequencing, assembly sequencing, and single-cell sequencing of germ cells.

[0043] Further, it is found that the same rule applies to positions such as chr4:226360:T:C:rs11248009 and chr4:228314:C:T:rs61794997.

[0044] Further, when a wgs sequencing is performed on a certain free blood type sample, if the following base count result is found, in general, it can be concluded that at the position of chr4:213231, although only 1X sequencing is reached, according to the equivalent SNP information of linkage, the equivalent depth of chr4:213231 can be obtained as 4X, and the ro:ao is 2:2.

[0045] chrom pos rsid ref alt dp ro ao chr4 213231 rs5003260 G A 1 1 0 chr4 221985 rs6830496 G A 1 0 1 chr4 226360 rs11248009 T C 1 1 0 chr4 228314 rs61794997 C T 1 0 1

[0046] Based on the above possible rules of genetic loci, the human genome is explored and traversed to find equivalent depth that can be improved by wgs sequencing. Various ways and algorithms can be used to obtain the above rules for application, such as using the genomic information in the public thousand human genome project to calculate correlation and linkage equilibrium strength. As shown in the following, each column represents the linkage strength of the SNP to chr4:213204:rs4624608, and the horizontal coordinate can be regarded as the chromosome, which is the distance in base pairs between chr4:213204:rs4624608. Different colors correspond to different LD values. Due to the distance distribution is far away, the column chart may be fused together. fig. 3

[0047] Limitations of equivalent depth

[0048] a. If the DNA content extracted from the evidence is limited, such as only 2 genomes of 1 cell, the equivalent depth may exceed 2X, but it cannot be judged as a repeated sequence, and the result can be used for typing, but cannot be used for screening small signals.

[0049]

[0050] ​b.According to the screening of the target population, there is a biological error in the data source itself, and prenatal non-invasive parentage identification can be assisted by maternal white blood cell data.

[0051] Further, the inventors have invented a low-depth WGS sequencing method (with a sequencing depth as low as 0.25X) that can accurately determine the fetal DNA concentration.

[0052] The present patent is based on the following principle: SNP (Single Nucleotide Polymorphism) is a common form of genetic variation, which exists widely in the human genome. Linkage disequilibrium (LD) is used to describe the non-random association between two or more genetic markers, usually SNPs. When the combination of alleles of certain SNPs appears more frequently in a population than expected from the individual allele frequencies, this phenomenon is called linkage disequilibrium. D is the basic unit of LD (linkage disequilibrium), which measures the deviation of the observed block frequency from the expected frequency under the equilibrium state. LD can be measured from 0-1, the higher the LD, the stronger the linkage between two sites if they are linked. The present patent takes the linkage SNP block as the calculation unit, which can equivalently increase the depth.

[0053] The method comprises the following steps:

[0054] S1: Low-depth WGS sequencing of maternal free DNA samples during pregnancy to obtain sequencing data R.

[0055] S2: According to the correspondence between the point and the linkage SNP block, screen the typing data Q of the corresponding sites constituting the linkage SNP block on the reference genome (specifically, the WGS data of the corresponding population in the 1KGP data) in R.

[0056] S3: Calculate the fetal concentration according to the typing data Q and take the linkage SNP block as the unit (use heterozygosity to calculate, specifically can use binomial distribution method).

[0057] The sequencing data is stored in BAM (Binary Alignment File) or VCF (Variant Call File) format, which is convenient for subsequent analysis.

[0058] Among them, steps S2-S3 can use python or R to parse the correspondence between the point and the linkage SNP block corresponding to the sequencing data. Data screening and typing can be performed through existing bioinformatics tools (such as PLINK, GATK, etc.); heterozygosity analysis and concentration calculation can be performed through statistical software packages (such as SciPy, R's stats package).

[0059] The correspondence between the point and the linkage SNP region block is obtained by screening a reference genome (specifically the same as the population to be detected) according to a predetermined rule. The linkage SNP region block of the present application is a linkage SNP region block composed of a plurality of linkage sites with high linkage degree. According to a specific algorithm, the linkage sites in the linkage SNP region block have a long distance and meet specific conditions, so that the sequencing depth of the linkage SNP region block is the superposition of the sequencing depth of the corresponding typing linkage sites in it when calculating the concentration, which simplifies the processing process of the sequencing data. The reference genome is the public WGS data of the corresponding population or the WGS data obtained by sequencing a large number of corresponding populations. The method of the present application can be applied to sexually reproducing diploid organisms. The present application is illustrated by humans (Chinese population), but is not limited thereto. The above method can be used for other animals (such as monkeys, mice, rats, rabbits, etc., mainly mammals), other populations (such as African populations, American populations, etc.). Specifically, the correspondence between the point and the linkage SNP region block constitutes a hap database. The linkage SNP region block is composed of a plurality (such as more than 10, preferably more than 30) of linkage sites with high linkage degree.

[0060] Further, in step S1, the low-depth WGS data after sequencing is processed, and other methods can also be used for processing. The processing process is as follows:

[0061] S101: The low-depth WGS data is processed by fastp software to remove the adapters on the reads.

[0062] S102: The processed data in step S101 is aligned and corrected with the reference genome GRCh38 by bwa software to obtain a first alignment file in bam format.

[0063] S103: The first alignment file corrected in step S102 is processed by picard or genecore software to remove duplicate reads, and a second alignment file not containing duplicate reads is obtained.

[0064] S104: The low-quality data is filtered by bcftools mpileup to obtain sequencing data R (specifically a site typing file VCF).

[0065] The predetermined rule is:

[0066] (1) From the G WGS data of the autosomes in the reference genome, the SNPs with a population frequency of 0.2-0.8 are taken and the sites and regions that do not meet the requirements are removed. G is greater than or equal to 5, and preferably G is the number of all autosomes corresponding to the population to be detected.

[0067] (2) On a chromosome, calculate the LD value between the first SNP site SNP1 and the subsequent SNP sites in sequence from front to back. In theory, the closer the points, the higher the degree of imbalance. The farther the points, the higher the probability of disruption by recombination, and the lower the degree of linkage. Statistics of these values until at least K LD values < first threshold T1 cutoff in the first M consecutive LD values, SNP1 to the cutoff site as the first region block. This patent only calculates the LD value on a single chromosome. Step (2) can reduce the amount of calculation.

[0068] (3) Screen the SNP sites in the first region block, and select the SNP sites with LD values greater than the second threshold T2 as linkage sites. The linkage sites in the first region block form the second region block. Then judge the region block corresponding to the second region block. If hap1 + hap2 > G * T3, the second region block is retained as a linkage SNP region block; otherwise, the second region block is deleted (then A does not have a linkage SNP region block); wherein hap1 and hap2 represent the largest two region blocks, and T3 represents the third threshold.

[0069] (4) In this way, the subsequent adjacent linkage sites (the previous SNP site after step (1) processing, and simultaneously satisfying the LD value greater than the second threshold T2) are fixed, and then the LD value is calculated with the subsequent SNP sites. According to steps (2) and (3), all linkage SNP region blocks on the chromosome are obtained.

[0070] (5) In this way, according to steps (2)-(4), the linkage SNP region blocks of multiple chromosomes are obtained; the number of linkage SNP region blocks is greater than 10,000, and the number of linkage sites in the linkage SNP region block is greater than 10.

[0071] Among them, removing sites and regions that do not meet the requirements include removing indel type sites, removing STR related SNP sites, removing MSI related SNP sites, removing Hardy-Weinberg imbalance points, removing regions with the possibility of duplication, and removing regions with high GC content and low complexity near SNP, etc. to reduce the amount of calculation and reduce the interference on the subsequent program. The above process is a conventional process, which is well known to those skilled in the art, and therefore the detailed description is omitted.

[0072] Specifically, M is greater than or equal to 10, M-K is 1-4, T1 is 0.01-0.20, T2 is 0.90-0.99, and T3 is 0.95-0.99. Among them, M, K, T1, T2 and T3 can be adjusted according to the purpose, precision or platform, etc. Preferably, M is 12, K is 10, T1 is 0.1, T2 is 0.95, and T3 is 0.98.

[0073] Specifically, taking the population of China as an example, the WGS data of 22 pairs of autosomes of CHS and CHB in the 1KGP are used, and of course, the WGS data of CDX can also be used or added for the population of China. If used for the population of other countries, the population data of the corresponding country is used, so this step does not limit the population. The number of SNPs of 22 pairs of chromosomes obtained from CHS and CHB is shown in Table 1:

[0074] Table 1 Number of SNPs of 22 pairs of chromosomes

[0075] chromosome chrom SNP num chr1 243117 chr2 248355 chr3 215927 chr4 231182 chr5 197186 chr6 211535 chr7 186055 chr8 173553 chr9 135534 chr10 157705 chr11 156941 chr12 147287 chr13 115835 chr14 102990 chr15 91245 chr16 97307 chr17 80846 chr18 91798 chr19 72402 chr20 70515 chr21 46598 chr223 43084

[0076] Specifically, the predetermined rules are:

[0077] (1) Take 416 WGS data of autosomes of CHS and CHB of two populations in the 1KGP (The 1000 Genomes Project, 1000G / 1KGP), take SNPs with a population frequency of 0.2-0.8, and remove indel type sites, remove SNP sites related to STR, remove SNP sites related to MSI, remove haplotype imbalance points, remove regions with possible duplication, and remove regions with high GC content and low complexity near SNPs.

[0078] (2) On a chromosome, calculate from front to back in order, fix the first SNP as SNP1, calculate the LD value between the subsequent SNP site and SNP1, and count these values until at least 10 LD values in the last 12 consecutive LD values are less than 0.1 cutoff, and SNP1 to the cutoff site as the first region block.

[0079] (3) Screen the SNP sites in the first region block, and take the SNP sites with an LD value greater than 0.95 as linkage sites, and the linkage sites in the first region block form the second region block; then judge the region block corresponding to the second region block, if hap1+hap2>416*0.98, the second region block is retained as a linkage SNP region block; otherwise, the second region block is deleted.

[0080] (4) In this way, the subsequent adjacent linkage sites are fixed, and the LD value is calculated with the subsequent SNP sites in turn, and all linkage SNP region blocks on the chromosome are obtained according to steps (2) and (3).

[0081] (5) In this way, the linkage SNP region blocks of 22 chromosomes are obtained according to steps (2)-(4) to constitute the hap database.

[0082] Preferably, in step (2), only the LD values of the SNP sites with abs(AF-0.5) less than a fourth threshold T4 are calculated to further reduce the number of LD values calculated, and T4 is 0.01-0.04. Specifically, T4 is 0.02.

[0083] In step S3, the linkage SNP region block is represented as a whole, and the linkage SNP region block is aligned with the reference genome, typed according to the sequencing depth ratio, and the sequencing depth of the linkage SNP region block is the superposition of the sequencing depth of the corresponding typed linkage sites therein. The fetal concentration is calculated by the binomial distribution method, and the reads of the linkage SNP region block consistent with the reference genome are denoted as nA, and the reads of the linkage SNP region block inconsistent with the reference genome are denoted as na, and n=nA+na; if af=nA / n<0.2 or >0.8, then it is considered that the reads with a proportion of <0.2 are provided by the fetus. The present application illustrates the calculation results of the af values in 5 different cases, as shown in Table 2:

[0084] Table 2 Calculation results of af values in 5 different cases

[0085]

[0086] As fig. 4 shown, plot n and af values, and in the typing diagram, each point represents a SNP region block site, the horizontal coordinate is the sequencing depth of the region block site, and the vertical coordinate represents the proportion of wild type of the SNP site. The heterozygosity distribution diagram of each linkage SNP region block detected can be obtained.

[0087] Specifically, the binomial distribution method is specifically:

[0088] S301: Obtain a site set X' in the typing data B that satisfies the following requirements,

[0089] X'={xi|0<nai(S) / ni(S)<0.2∪0<nAi(S) / ni(S)<0.2}.

[0090] S302: Calculate the probability P of a certain site on the site set X' according to formula I,

[0091]

[0092] Wherein, Pm=0.4.

[0093] S303: Obtain pmax when the cumulative probability of all P of the site set X' is maximum by using the maximum likelihood method.

[0094] S304: Fetal DNA concentration N=2pmax.

[0095] Advantages

[0096] (1) In traditional methods, some cffDNA concentration calculation methods based on Y chromosome markers are only applicable to male fetuses. The present application calculates cffDNA concentration directly through the heterozygosity of linkage SNP region blocks based on low-depth WGS data, which is not affected by the gender of the fetus in NIPT practical applications.

[0097] (2) Many traditional cffDNA concentration calculation methods require a father sample as a reference to distinguish between maternal and fetal genetic information. Through linkage SNP region block heterozygosity analysis, fetal DNA information can be directly extracted from maternal blood during pregnancy without the need for a father sample and without the need for additional experiments.

[0098] (3) Low-depth whole genome sequencing (WGS) has lower sequencing depth and less data, so the sequencing cost is significantly lower than high-depth sequencing.

[0099] (4) High stability: Linkage SNP region blocks have high linkage disequilibrium (LD) and strong genetic stability, and are less affected by random variation. Calculating in units of linkage SNP region blocks can effectively reduce random errors and improve the stability and reliability of the results.

[0100] (5) Superior performance: Through linkage SNP region block heterozygosity analysis, this method can more accurately calculate cffDNA concentration, especially in low concentration situations, while still maintaining high sensitivity and specificity. Since linkage SNP region blocks have a wide coverage range, the present application can fully utilize genomic information to improve detection performance. Therefore, by optimizing the algorithm and analyzing linkage SNP region blocks, the present application can accurately calculate cffDNA concentration in low-depth WGS data, reducing costs while maintaining high accuracy.

[0101] (6) This method directly uses low-depth WGS sequencing data without the need for additional experimental steps (such as targeted sequencing or PCR amplification), simplifying the experimental process and shortening the detection time. The efficiency of this method makes it more suitable for large-scale clinical applications.

[0102] (7) Linkage SNP region block data can be obtained from public genome databases or high-depth WGS data, making it suitable for different populations. This versatility makes the present application more flexible in different regions and populations.

[0103] (8) The core principle of the method (analysis based on linkage SNP region block) is not only suitable for cffDNA concentration calculation, but also can be extended to other fields such as circulating tumor DNA (ctDNA) analysis in liquid biopsy of tumor. This scalability provides more possibilities for future research and application. BRIEF DESCRIPTION OF DRAWINGS

[0104] fig. 1 is a flowchart of the method for evaluating the concentration of fetal DNA in maternal free blood during pregnancy based on low-depth WGS provided by the embodiments of the present application;

[0105] fig. 2 is a flowchart of processing machine data under low-depth WGS;

[0106] fig. 3 is a control relationship diagram of base position and LD value of chr4:213204 site;

[0107] fig. 4 is a heterozygosity distribution diagram of linkage SNP region block;

[0108] fig. 5 is a correlation diagram of fetal concentration obtained by low-depth WGS technology and NGS technology. DETAILED DESCRIPTION

[0109] In order to make the purpose, technical scheme and advantages of the present application clearer, the present application will be further described in detail below with reference to the drawings.

[0110] Example 1

[0111] The process of obtaining the correspondence between the site and the linkage SNP region block is as follows:

[0112] (1) Take 416 WGS data of autosomes (22 pairs) of CHS (Han Chinese South) and CHB (Han Chinese in Beijing, China) two populations in 1KGP (The 1000 Genomes Project, 1000G / 1KGP), and take SNP with a population frequency of 0.2-0.8. Remove indel type sites, remove STR related SNP sites, remove MSI related SNP sites, remove regions with possible duplication, and remove regions with high GC content and low complexity near SNP.

[0113] (2) On a chromosome, calculate the LD value between the first SNP (SNP1) and the subsequent SNP site (SNP site with abs(AF-0.5) less than 0.02) from front to back in sequence, and count these values until at least 10 of the 12 consecutive LD values are <0.1 cutoff. SNP1 to the cutoff site is the first regional block.

[0114] Wherein, the calculation process of LD value is consistent with the prior art. The LD value calculation method is shown as follows; for example, the SNP genotype at a certain position is T, and the allele genotype is C; and the SNP genotype at a certain position connected thereto is T, and the allele genotype is C. Then there are TT, TC, CT and CC four regional blocks in the population, and sum_ = sum([TT, TC, CT, CC]), PT1 = (TT + TC) / sum_, PT2 = (TT + CT) / sum_, D = (TT / sum_) - PT1 * PT2, LD = D * D / (PT1 * PT2 * (1-PT1) * (1-PT2)). Taking the site chr4:213204 as an example, the LD value of the site is shown as follows: fig. 3

[0115] (3) The SNP sites in the first regional block are screened, and the SNP sites with LD value greater than 0.95 are taken as linkage sites. The linkage sites in the first regional block constitute the second regional block. The regional block corresponding to the second regional block is judged again. If hap1 + hap2 > 416 * 0.98 = 408, the second regional block is retained as a linkage SNP regional block; otherwise, the second regional block is deleted. Taking the site chr4:213204 as an example,

[0116]

[0117]

[0118]

[0119] hap1 + hap2 = 365 + 47 = 412 > 408, so the second regional block corresponding to the site 213204 is retained as a linkage SNP regional block.

[0120] (4) Similarly, the subsequent adjacent linkage sites are fixed, and the LD value is calculated with the subsequent SNP sites in turn, and all linkage SNP regional blocks on the chromosome are obtained according to steps (2) and (3).

[0121] (5) Similarly, the linkage SNP regional blocks of 22 chromosomes are obtained according to steps (2)-(4) to constitute the hap database. ​

[0122] Example 2

[0123] Based on the low-depth WGS evaluation of the concentration of fetal DNA in the maternal free blood during pregnancy, the following steps are included:

[0124] S1: low-depth WGS sequencing is performed on the maternal free DNA sample during pregnancy to obtain sequencing data R (which can be stored in fastp format, for example).

[0125] S2: The typing data Q of the corresponding sites constituting the linkage SNP region block on the reference genome in R is screened according to the correspondence between the sites obtained in Example 1 and the linkage SNP region block (specifically, the hap database constructed by the method of Example 1) (see Table 2).

[0126] S3: fetal concentration is calculated according to the typing data Q and in units of linkage SNP region blocks; step S3 specifically includes:

[0127] S301: obtain a site set X' in the typing data Q that meets the following requirements,

[0128] X' = {x i |0 < na i (S) / n i (S) < 0.2 U 0 < nA i (S) / n i (S) < 0.2}.

[0129] The reads of the linkage SNP region block consistent with the reference genome are denoted as nA, and the reads of the linkage SNP region block inconsistent with the reference genome are denoted as na, n = nA + na.

[0130] S302: the probability P of a site in the site set X' is calculated according to formula I,

[0131]

[0132] wherein, Pm = 0.4.

[0133] S303: the maximum likelihood method is used to obtain pmax when the cumulative probability of all P of the site set X' is maximum.

[0134] S304: fetal DNA concentration N = 2pmax.

[0135] The detection results are as follows

[0136] Example 3

[0137] The method of Example 2 was used to perform PE150 low-depth WGS sequencing on the G99 sequencing platform for 152 maternal plasma samples collected during pregnancy (Lanhai Bio). The fetal concentration is denoted as ngs-wgs.

[0138] The binomial distribution method was used for calculation, and NGS data was obtained by high-throughput liquid-phase hybrid capture technology and sequencing by MGISEQ-2000. The fetal concentration is denoted as ngs-hybrid.

[0139] The data of the samples (ngs-wgs range: 0.01-0.16) are shown in Table 3 below

[0140] Table 3 Data of samples (ngs-wgs range: 0.01-0.05)

[0141]

[0142]

[0143]

[0144]

[0145] Wherein, ngs-wgs represents the concentration obtained by WGS data, and ngs-hybrid represents the concentration obtained by the existing hybrid method.

[0146] The correlation coefficient of the two can be obtained by plotting ngs-wgs as the Y axis and ngs-hybrid as the X axis, as shown in fig. 5 r2=0.981. As can be seen from fig. 5 , the fetal concentrations obtained by the two methods have a correlation of more than 0.98, and the difference is very small. In addition, as can be seen from fig. 5 , there are a large number of low-concentration samples in the 152 samples.

[0147] The above only describes the preferred embodiments of the present application and is not intended to limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method for assessing the concentration of fetal DNA in maternal plasma during pregnancy by linkage of blocks of polymorphic loci, characterized in that, The method comprises: S1: establishing a linkage relationship database of polymorphic sites; S2: performing low-depth WGS sequencing on the sample to obtain sequencing data; S3: screening typing data of corresponding sites constituting a linkage polymorphic site region block on a reference genome in the sequencing data according to the linkage relationship of the polymorphic sites; S4: performing fetal concentration calculation based on a binomial distribution according to the typing data and in units of linkage polymorphic site region blocks; wherein the linkage polymorphic site region block is a linkage polymorphic site region block composed of multiple linkage sites with high linkage degrees, containing one core site and one to multiple other sites linked thereto, and the reference genome is public genomic WGS data of a corresponding population or population; uniting in linkage polymorphic site region blocks refers to aligning the linkage polymorphic site region block as a whole with the reference genome, typing according to the sequencing depth ratio, and the sequencing depth of the linkage polymorphic site region block being the superposition of the sequencing depths of the corresponding typed linkage sites therein.

2. The method of claim 1, wherein, In step S1, the linkage relationship of the polymorphic sites is obtained by at least one of the following ways: (1) obtained by processing public genomic WGS data; (2) obtained by processing a large amount of high-depth WGS measured data; (3) obtained from haplotype cell sequencing data; (4) obtained by processing long sequencing mode data; (5) obtained by querying public databases or literature.

3. The method of claim 2, wherein the linkage relationship of the polymorphic sites is obtained by processing public genomic WGS data or by processing a large amount of high-depth WGS measured data.

4. The method of claim 1, wherein steps S2-S4 are specifically: S2: performing low-depth WGS sequencing on a maternal free DNA sample during pregnancy to obtain sequencing data; S3: screening typing data of corresponding sites constituting a linkage SNP region block on a reference genome in the sequencing data according to the correspondence between the sites and the linkage SNP region block; S4: calculating fetal concentration in units of linkage SNP region blocks according to the typing data; wherein the correspondence between the sites and the linkage SNP region block is obtained by screening the reference genome, and the linkage SNP region block is a linkage SNP region block composed of multiple SNPs with high linkage degrees, containing one core site and one to multiple other sites linked thereto; the reference genome is public genomic WGS data of a corresponding population or population.

5. The method of claim 4, wherein, In step S2, the low-depth WGS sequencing data is processed after sequencing, and the processing process is as follows: S101: removing the adapters on the reads from the low-depth WGS sequencing data; S102: aligning and correcting the sequencing data processed in step S101 with the reference genome to obtain a first alignment file; S103: removing duplicate reads from the first alignment file corrected in step S102 to obtain a second alignment file not containing duplicate reads; S104: filtering low-quality data to obtain the sequencing data.

6. The method of claim 5, wherein the removing the adapters on the reads is performed using fastp software; wherein the raw data is aligned to the reference genome GRCh38 using bwa software to correct the alignment and obtain the bam format alignment file; wherein the duplicated reads are removed using picard or genecore software; and wherein the low quality data is filtered using bcftools mpileup.

7. The method of claim 3, wherein the screening is performed according to the following rules: (1) obtaining SNPs with a certain population frequency from autosomes in G WGS data of the reference genome and removing sites and regions that do not meet the requirements; (2) sequentially calculating from front to back on a chromosome, fixing the first SNP site as SNP1, calculating the LD value between the subsequent SNP site and SNP1, and counting the values until at least K LD values in the continuous M LD values are less than the first threshold T1, and the SNP1 to the cutoff site is taken as the first region block; hap1 and hap2 are the two most frequent heterozygous region blocks in the second region, P(hap1) is the frequency of the region block hap1 appearing in G data, and T3 is the third threshold; (3) fixing the subsequent adjacent linkage sites in turn, and then sequentially calculating the LD value with the subsequent SNP site to obtain all linkage SNP region blocks on the chromosome according to steps (2) and (3); and (4) similarly, obtaining linkage SNP region blocks of multiple chromosomes according to steps (2)-(4).

8. The method of claim 7, wherein the certain population frequency is between 0.2-0.

8.

9. The method of claim 7, wherein the number of linkage SNP region blocks is greater than 10,000, and the number of linkage sites in the linkage SNP region block is greater than 10. (3) screening the SNP sites in the first region block, taking the SNP sites with the LD value greater than a second threshold T2 as linkage sites, and the linkage sites in the first region block forming a second region block; then judging the region block corresponding to the second region block, if P(hap1)+P(hap2)>T3, the second region block is retained; otherwise, the second region block is deleted; wherein, The removing sites and regions that do not meet the requirements includes at least one of the following: removing indel type sites, removing SNP sites related to STR, removing SNP sites related to MSI, removing Hardy-Weinberg disequilibrium points, removing regions with possible duplication, and removing regions with high GC content and low complexity near the SNP. M is greater than or equal to 10, M-K is 1-4, T1 is 0.01-0.20, T2 is 0.90-0.99, and T3 is 0.95-0.

99. In step (1), 22 pairs of autosomal WGS data of two populations CHS and CHB in 1KGP are used, and G is 416.

13. The method of claim 7, wherein, 416 autosomal WGS data of two populations CHS and CHB in 1KGP are taken, the population frequency is between 0.2-0.8, M is 12, K is 10, T1 is 0.1, T2 is 0.95, and T3 is 0.

98.

10. The method of claim 7, wherein, In step (2), only the LD value of the SNP site with an absolute value of (AF-0.5) less than the fourth threshold T4 is calculated, and the T4 is 0.01-0.

04.

11. The method of claim 7, wherein, ​ 12. The method of claim 7, wherein, ​ ​ ​ 14. The method according to claim 7 or 13, characterized in that, ​ 15. The method of claim 4, wherein fetal concentration is calculated based on binomial distribution, reads that match the linkage SNP region block to the reference genome are denoted as nA, reads that do not match the linkage SNP region block to the reference genome are denoted as na, n = nA + na; if nA / n < 0.2 or > 0.8, then reads that account for < 0.2 are considered to be provided by the fetus.

16. The method of claim 15, wherein the specific steps comprise S301: obtaining a set of sites X' from the typing data that satisfy the following requirements; X' = {x i |0 < na i (S) / n i (S) < 0.2 U 0 < nA i (S) / n i (S) < 0.2} S302: calculating the probability P of a site in the set of sites X' according to the following formula I, ； wherein 0 < p < 0.5, Pm = 0.4; S303: obtaining pmax when the cumulative probability of all P of the set of sites X' is maximum by using the maximum likelihood method; S304: fetal DNA concentration N = 2pmax.

17. The method of claim 3, wherein the gamete obtained region block of a specific population is screened according to the following rules: obtaining the gamete obtained region block of a specific population by selecting the semen of 400-1000 individuals in the population, isolating and obtaining single sperm, performing single-cell next-generation sequencing to obtain sequencing data; aligning the sequencing data to the reference genome to obtain 400-1000 haplotype data; obtaining the frequency data of independent polymorphisms of the population by aligning the 400-1000 haplotype data; the germ cell haplotype is a kind of specialization for describing the linkage region block, the SNP with a frequency of 0.2-0.8 in the population is taken, and the sites and regions that do not meet the requirements are removed; taking each SNP above as a starting point, expanding all SNPs within 100 kbp to form a linkage group for each SNP, and screening the SNP linkage group by the following conditions: the sum of the frequencies of a specific region block P in the linkage group and its allotropic region block Q is greater than 0.95, and the specific region block P covers more than N SNPs, wherein N is an integer greater than or equal to 2; the specific region block P includes a deletion region block; if a region block is AAAaaa, then aaaAAA is its allotropic region block, and A-Aaaa and AAA--a are its deletion region block.

18. The method of claim 3, wherein the screening is performed according to the following rules: selecting 400-1000 individuals in a population, obtaining whole genome haplotype sequencing data with fragment length >50kb by third-generation sequencing, obtaining 400-1000 haplotype data by aligning the sequencing data to a reference genome, obtaining frequency data of independent polymorphisms in the population by aligning the 400-1000 haplotype data, long sequencing haplotype is a kind of specialization for describing linkage block, taking SNPs with a frequency of 0.2-0.8 in the population and removing sites and regions that do not meet the requirements, expanding all SNPs within 100kbp connected to each SNP above as a linkage group, and screening the SNP linkage group by the following conditions: the sum of the frequency of a certain specific block P and its alien block Q in the linkage group is greater than 0.95, and the block P covers more than N SNPs, wherein N is an integer greater than or equal to 2; the specific block P includes a missing block; if a certain block is AAAaaa, then aaaAAA is its alien block, and A-Aaaa and AAA--a are its missing blocks.

19. The method of claim 3, wherein the screening is performed according to the following rules: using the public database LDLink, screening the block of a specific population by LDhap of the LDLinkR software package, taking SNPs with a frequency of 0.2-0.8 in the database of Asian population and removing sites and regions that do not meet the requirements; expanding all SNPs within 100kbp connected to each SNP above as a linkage group, and screening the SNP linkage group by the following conditions: the sum of the frequency of a certain specific block P and its alien block Q in the linkage group is greater than 0.95, and the specific block P covers more than N SNPs, wherein N is an integer greater than or equal to 2; the specific block P includes a missing block; if a certain block is AAAaaa, then aaaAAA is its alien block, and A-Aaaa and AAA--a are its missing blocks.

20. The method of claim 4, wherein, The linkage SNP block is represented as a whole by aligning the linkage SNP block to the reference genome, and is typed according to the sequencing depth ratio, and the sequencing depth of the linkage SNP block is the superposition of the sequencing depth of the corresponding linkage site.

21. The method of any one of claims 17-19, wherein N is an integer of 10-30.

22. The method of claim 2, wherein the haplotype cell is a germ cell; and the long sequencing mode is Sanger or third-generation sequencing.

Images