Method for detecting sample cross-contamination based on SNP sites and application thereof

Abstract

The application discloses a method for detecting sample cross contamination based on SNP sites and application thereof. The method comprises the following steps: screening SNP sites, simulating data, constructing a model, and evaluating contamination degree according to the model. The screening of SNP sites comprises the following steps: (1) downloading sequencing-related FASTQ files or bam files from a database; (2) performing variation detection on high-throughput sequencing data of each sample in the files to obtain gvcf format files of variation information of the samples; (3) merging all gvcf files of the samples to obtain vcf format variation information of all the samples; and (4) screening SNP sites according to heterozygous state population frequency and variation allele frequency. The application simulates data of stable SNP sites carrying high population frequency, constructs a model, and evaluates contamination degree according to the model. The AF values of these SNP sites in different samples are relatively stable, and are concentratedly distributed near 0, 0.5 and 1; and the dose change of the AF values after the SNP sites contaminate other samples can accurately reflect the degree of sample cross contamination, and the operation is simple, fast and convenient.

Description

Technical Field

[0001] This invention belongs to the field of biotechnology and relates to a method for detecting cross-contamination of samples based on SNP sites and its application. Background Technology

[0002] High-throughput sequencing technology can sequence hundreds or thousands of genes or the entire genome in a short time. It is characterized by high efficiency, low cost, and high specificity, and is widely used in clinical medical research, playing a crucial role in tumor diagnosis and treatment, genetic disease diagnosis, non-invasive prenatal testing, prenatal screening and diagnosis, and preimplantation genetic diagnosis and screening for in vitro fertilization (IVF). This technology involves numerous experimental steps. First, DNA / RNA is extracted from the sample, and the extracted DNA / RNA undergoes quality control. Then, the qualified molecules are fragmented, end-repaired with polyA tails and adapters, PCR amplification is performed, sequencing libraries are prepared, and the libraries undergo quality control. Finally, sequencing is performed to obtain sequencing data. Throughout the experimental process, cross-contamination between samples from the same batch may occur due to improper operation by researchers, or contamination from different species introduced by the laboratory environment. Furthermore, when multiple samples are mixed for sequencing, index tags are added to ensure a one-to-one correspondence between the obtained data and the sample. However, when samples from the same batch are sequenced together, adapter skipping can occur, causing sequencing data originally belonging to one sample to be incorrectly split into another, resulting in sample cross-contamination. Sample contamination has a significant impact on experimental results and can cause false positives, which should be avoided at all costs in clinical trials. Experimentally, the probability of adapter skipping is much lower with dual adapters than with single adapters; therefore, dual adapter library construction can be used during sequencing to avoid this problem. However, contamination may be unavoidable due to uncontrollable factors. In such cases, contamination should be detected promptly and remedial measures taken.

[0003] Currently, it is difficult to detect contamination through experiments, and some bioinformatics methods are needed to solve this problem. Sequencing contamination mainly comes from interspecies contamination, intrasample contamination and intersample contamination, and different contaminations require different solutions. (1) Interspecies contamination refers to the presence of DNA or RNA from other species in the sample, which may be due to the presence of DNA or RNA from other species in the laboratory environment. This can be eliminated by comparing the sequencing fragments to the reference genome and calculating the alignment rate. The human reference genome is relatively complete, and the alignment rate can generally reach more than 99%. If the alignment rate is too low, it is very likely that other species were contaminated during the sampling or library construction process. It is worth noting that the default parameters of the bwamem algorithm (parameter: -k: minimum seed length [default 19]) are relatively loose. When contamination occurs in species with similar genomes, the alignment rate is also high. In this case, the -k parameter can be appropriately increased or other algorithms can be used to assess whether the sample is contaminated by other species. (2) Intrasample contamination refers to the presence of different types of DNA or RNA in the same sample, which may be due to improper operation during sample collection, extraction, storage and other processes. The most common issue is impure cancer tissue sampling, where normal tissue is mixed in, leading to decreased sensitivity. Alternatively, impure adjacent tissue sampling, where cancer tissue is mixed in, can prevent the detection of somatic mutations. A commonly used assessment method is Sequenza. (3) Inter-sample contamination refers to the mixing of tissues from other samples, resulting in a large number of false-positive mutation sites. This type of cross-contamination is difficult to detect. Kristian Cibulskis and Aaron McKenna developed an algorithm called ContEst, which uses Bayesian methods to calculate the posterior probability of contamination based on population mutation frequency information. It is mainly used to check whether sample contamination occurs between cancer and adjacent tissues, and whether they are cancer and adjacent tissue samples from the same patient. It is not suitable for assessing single-sample cross-contamination.

[0004] In summary, due to the complexity of the genome, the frequencies of variant alleles at the mutation sites in the wild-type, heterozygous, and homozygous states fluctuate greatly, making accurate detection of cross-contamination between samples impossible. Therefore, developing a method to detect cross-contamination, improve the stability of variant allele frequencies, and enhance the accuracy of cross-contamination detection has become one of the urgent problems to be solved in the field of biotechnology. Summary of the Invention

[0005] To address the shortcomings of existing technologies and practical needs, this invention provides a method for detecting sample cross-contamination based on SNP sites and its application. The method involves screening stable SNP site simulation data with high population frequency to construct a model, and then assessing the degree of contamination based on the model. The AF values ​​of these SNP sites are relatively stable across different samples, concentrated around 0, 0.5, and 1. The dose change of their AF values ​​after contaminating other samples can accurately reflect the degree of sample cross-contamination. The method is simple, fast, and convenient.

[0006] To achieve this objective, the present invention adopts the following technical solution:

[0007] In a first aspect, the present invention provides a method for detecting cross-contamination of samples based on SNP sites, the method comprising: screening SNP sites, simulating data, constructing a model, and assessing the degree of contamination based on the model;

[0008] The screening of SNP sites includes the following steps:

[0009] (1) Download the sequencing-related FASTQ or bam files from the database;

[0010] (2) Perform mutation detection on the high-throughput sequencing data of each sample in the file to obtain the mutation information of the sample in gvcf format file;

[0011] (3) Merge all sample gvcf files to obtain vcf format variation information for all samples;

[0012] (4) Screen SNP sites based on the frequency of heterozygous populations and the frequency of variant alleles.

[0013] This invention constructs a model by screening stable SNP loci simulation data that carry high population frequency, and assesses the degree of contamination based on the model. The AF values ​​in different samples are relatively stable and concentrated around 0, 0.5 and 1, which can accurately reflect the degree of cross-contamination of samples. The operation is simple, fast and convenient.

[0014] Allele frequency (AF) refers to the frequency of a particular genotype variation in a sample. AF values ​​can roughly infer the genotype status: wild-type, heterozygous, or homozygous. Theoretically, the allele frequency for wild-type variants is around 0, for heterozygous around 0.5, and for homozygous around 1. If contamination from other samples is present, the AF values ​​for these three states will fluctuate. The change in the 2*AF value after contamination can, to some extent, represent the degree of contamination. However, due to the complexity of the genome, including numerous repetitive sequences, high heterozygosity, abnormal GC content, capture bias, and systematic sequencing errors, the AF values ​​for variant sites in all three states will fluctuate, especially for heterozygous and homozygous states where the fluctuations are very large. AF values ​​were statistically analyzed for the detected original mutations. Heterozygous SNPs had AF values ​​ranging from as low as 0.05 to as high as 0.85, concentrated around 0.5 with fluctuations reaching 0.4. Therefore, it is difficult to assess contamination levels using heterozygous SNPs in the test sample. Homozygous SNPs had AF values ​​around 1, with only a very small number having AF values ​​between 0.75 and 1. Therefore, variations in the AF values ​​of homozygous SNPs in the test sample can reflect the degree of contamination to some extent. Wild-type homozygous SNPs had an AF value of 0, indicating the absence of background interference and better reflecting the degree of contamination. Therefore, in test samples, wild-type SNPs should be prioritized for contamination assessment, followed by homozygous SNPs; heterozygous SNPs are not recommended.

[0015] Preferably, the sequencing in step (1) includes any one of whole genome sequencing, whole exome sequencing, or target sequence capture high-throughput sequencing.

[0016] Preferably, the SNP site screening in step (4) includes: retaining the target variant, deleting INDEL sites, retaining SNP sites whose FILTER column is PASS and which meet the requirements of heterozygous population frequency and untyped proportion, calculating the standard deviation of the AF value of each SNP site in the heterozygous state and the standard deviation of the AF value in the homozygous state in the normal population, summing them, and finally sorting them according to the SD value from smallest to largest, selecting sites with an AF value between 0.4-0.6 and 0.9-1 with a proportion greater than 0.30. The smaller the SD value, the more concentrated and stable the AF value distribution of this SNP site in the database sample. The ultimate goal is to select SNP sites with a relatively concentrated AF value distribution, mainly around 0, 0.5, and 1, to minimize the impact caused by the fluctuation of the AF value of the site itself.

[0017] The point values ​​in the range of 0.4-0.6 can specifically be 0.4, 0.42, 0.44, 0.46, 0.5, 0.52, 0.54, 0.56, 0.58, 0.6, etc.

[0018] The point values ​​in the above 0.9-1 can be 0.9, 0.91, 0.92, 0.93, 0.94, 0.95, 0.96, 0.97, 0.98, 1, etc.

[0019] Preferably, the target mutation includes conventional all external mutations.

[0020] Preferably, the frequency of the heterozygous population is 0.3-0.7.

[0021] The point values ​​in the range of 0.3-0.7 can specifically be 0.3, 0.36, 0.4, 0.44, 0.46, 0.5, 0.6, 0.67, 0.68, 0.68, 0.7, etc.

[0022] Preferably, the unclassified proportion is <0.10%.

[0023] Preferably, the number of sites is 200-2000.

[0024] The point values ​​in the range of 200-2000 can specifically be 200, 400, 600, 800, 1000, 1400, 1600, 1800, 1900, 2000, etc.

[0025] Preferably, the frequency of the variant alleles in step (4) is 0.05-1.

[0026] The point values ​​in the above 0.05-1 can be 0.05, 0.06, 0.07, 0.08, 0.09, 0.1, 0.2, 0.3, 0.6, 0.8, 1, etc.

[0027] Preferably, the simulated data includes: aligning sample mixed reads to a reference genome according to an algorithm, generating alignment files in BAM or SAM format, and then reordering and removing duplicate reads from the alignment files.

[0028] Preferably, the algorithm includes the bwa mem algorithm.

[0029] Preferably, the method for constructing the model includes univariate linear regression or Loess regression.

[0030] Preferably, the univariate linear regression includes: calculating the dose change of AF value to 0, 0.5 and 1 for each SNP site, and retaining the site closest to 0.

[0031] Preferably, the formula for the model is:

[0032] y = αx + β

[0033] Where y is the actual simulated pollution proportion, x is the degree of pollution, calculated using AF×2, α is the regression coefficient, and β is the intercept, calculated from the simulation data.

[0034] Secondly, the present invention provides a SNP site for detecting cross-contamination of samples, wherein the SNP site includes chr1: 1875858C->G, chr1: 11561593G->A, chr2: 272203C->T, chr2: 3392295A->G, chr3: 5024771T->C, chr3: 6903297C->T, chr4: 983060T->C, chr4: 983809C->T, chr5: 174106G->A, chr5: 1111 Any one or at least two of the following: 7583C->T, chr6:1313470C->T, chr6:11185533G->A, chr7:1482009G->C, chr7:11509561G->C, chr8:1616718A->G, chr8:10468172G->A, chr9:286593C->A, chr9:21350620G->A, chr10:3202140C->T, chr11:280816A->T.

[0035] Thirdly, the present invention provides the application of the method described in the first aspect in detecting cross-contamination of samples.

[0036] Compared with the prior art, the present invention has the following beneficial effects:

[0037] This invention constructs a model by screening stable SNP loci simulation data that carry high population frequency, and assesses the degree of contamination based on the model. The AF values ​​in different samples are relatively stable and concentrated around 0, 0.5 and 1, which can accurately reflect the degree of cross-contamination of samples. The operation is simple, fast and convenient. Attached Figure Description

[0038] Figure 1 Here is a flowchart for assessing cross-contamination of samples;

[0039] Figure 2 A distribution of AF values ​​for the selected SNP sites in the whole-external dataset;

[0040] Figure 3 This is a graph of a linear regression model;

[0041] Figure 4 This is a graph showing the contamination assessment results for normal samples.

[0042] Figure 5 A graph showing the pollution assessment results for a sample with a 1% pollution percentage;

[0043] Figure 6 A graph showing the pollution assessment results for a sample with a 2% pollution percentage;

[0044] Figure 7 A graph showing the pollution assessment results for a sample with a 3% pollution percentage;

[0045] Figure 8 A graph showing the pollution assessment results for a sample with a 4% pollution rate;

[0046] Figure 9 A graph showing the pollution assessment results for a sample with a 5% pollution rate;

[0047] Figure 10 A graph showing the pollution assessment results for a sample with a 10% pollution rate;

[0048] Figure 11 A graph showing the pollution assessment results for a sample with a 15% pollution rate;

[0049] Figure 12 A graph showing the pollution assessment results for a sample with a 20% pollution rate;

[0050] Figure 13 A graph showing the pollution assessment results for a sample with a 25% pollution rate;

[0051] Figure 14 A graph showing the pollution assessment results for a sample with a 30% pollution rate;

[0052] Figure 15 A graph showing the pollution assessment results for a sample with a 35% pollution rate;

[0053] Figure 16 A graph showing the pollution assessment results for a sample with a 40% pollution rate;

[0054] Figure 17 A graph showing the pollution assessment results for a sample with a 45% pollution rate;

[0055] Figure 18 The graph shows the pollution assessment results for a sample with a 50% pollution rate. Detailed Implementation

[0056] To further illustrate the technical means and effects of this invention, the following description, in conjunction with embodiments and accompanying drawings, provides a further explanation of the invention. It is understood that the specific embodiments described herein are merely illustrative of the invention and not intended to limit it.

[0057] Where specific techniques or conditions are not specified in the examples, they shall be performed in accordance with the techniques or conditions described in the literature in this field, or in accordance with the product instructions. Reagents or instruments whose manufacturers are not specified are all conventional products that can be purchased through legitimate channels.

[0058] To facilitate a clearer understanding of this invention, certain terms are first defined. As used herein, unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. All patents and publications mentioned herein are incorporated herein by reference in their entirety.

[0059] The term "Allele Frequency (AF)" refers to the frequency of variant alleles, describing the proportion of reads supporting the mutated base to the total number of reads at that locus. It is calculated as AD (allele depth) / DP (total depth).

[0060] Example 1

[0061] Screening for SNP sites.

[0062] (1) Download FASTQ or BAM files of full exotropy correlation from public databases such as 1000 Genomes, EBI and NCBI. The more samples, the better.

[0063] (2) For each sample’s high-throughput sequencing data, mutation detection is performed first to obtain the sample’s mutation information file (gvcf format).

[0064] (3) All samples are merged into gvcf files to obtain the variation information (vcf format) of all samples.

[0065] (4) Screen 2000 sites with a heterozygous population frequency between 0.3 and 0.7 and an AF value (number of reads supporting the variant base / total number of reads at the site) that is stable between 0.5 and 1. The specific screening method is as follows: Due to non-specific capture, the detected variants include many low-quality variants with low sequencing depth. These variants are likely to be false positives. Therefore, only target variants in the conventional whole exosome or other capture regions are retained, INDEL sites are deleted, and only SNP sites are retained. Then, SNP sites with a FILTER column as PASS, a heterozygous population frequency between 0.3 and 0.7, and an ungenerated proportion <0.10 are retained. The standard deviation of the AF value of each SNP site in the heterozygous state and the standard deviation of the AF value in the homozygous state in the normal population are calculated and summed. Finally, the sites are sorted from smallest to largest according to the SD value. The proportion of AF values ​​between [0.4, 0.6] and [0.9, 1] is greater than 0.3, and the top 2000 sites are selected. The goal is to select SNP loci with relatively concentrated AF values, particularly those around 0, 0.5, and 1. The first 20 selected loci are as follows: Figure 1 As shown, the statistical indicators for the screened SNP loci include Chrom: the chromosome number of the locus; Start: the start position of the locus; End: the end position of the locus; Ref: the reference base of the locus; Alt: the mutant base of the locus; Un_ratio: the population frequency of the locus that was not successfully genotyped; Wild_ratio: the population frequency of the wild-type homozygous state of the locus; Het_ratio: the population frequency of the heterozygous state of the locus; Hom_ratio: the population frequency of the homozygous state of the locus; SD: standard deviation.

[0066] Example 2

[0067] This embodiment involves model construction.

[0068] The simulated data samples were simulated with contamination rates of 1% (1M and 99M), 2% (2M and 98M), 3% (3M and 97M), 4% (4M and 96M), 5% (5M and 95M), 10% (10M and 90M), 15% (15M and 85M), 20% (20M and 80M), 25% (25M and 75M), 30% (30M and 70M), 35% (35M and 65M), 40% (40M and 60M), 45% (45M and 55M), and 50% (50M and 50M). First, the BWA MEM algorithm was used to align the mixed reads from two samples to the reference genome, generating alignment files in BAM / SAM format. Then, the alignment files were reordered and duplicate reads were removed.

[0069] Building Model

[0070] The AF values ​​of mutated bases at the selected SNP sites in the aligned bam / sam files were statistically analyzed. Theoretically, each SNP site has three states: wild-type homozygous (AF close to 0), heterozygous (AF close to 0.5), and homozygous (AF close to 1). If contamination from other samples is present, the AF values ​​for these three states will fluctuate, but the AF values ​​for heterozygous and homozygous states fluctuate more significantly, while the AF value for wild-type homozygous is close to 0. The 2*AF value after contamination represents the degree of contamination. Therefore, 2000 SNP sites that are wild-type homozygous in the predicted samples were selected for contamination assessment. The specific operation is as follows: the dose change of AF value to 0, 0.5, and 1 for each SNP site was calculated, retaining the site closest to 0. The median AF value of all wild-type homozygous SNPs was calculated, and AF*2 represents the degree of contamination (contamination_ratio). AF*2 was used as the independent variable x, and the actual simulated contamination ratio was used as the dependent variable y. A univariate linear regression model was constructed. The formula for the univariate linear regression equation is as follows:

[0071] y = αx + β

[0072] Where x is the independent variable, y is the dependent variable, α is the regression coefficient, and β is the intercept. Based on the simulation data, the regression coefficient is determined to be 1.135, and the intercept is -1.298. Once the regression coefficient and intercept are determined, the estimated value of y can be obtained from the observed values ​​of x. Figure 3 The graph shows the constructed linear regression model. The results indicate that the model fits the line well, with most points falling on the fitted line.

[0073] Example 3

[0074] This embodiment uses the model constructed in Embodiment 2 to assess the contamination level of experimental simulation data.

[0075] Samples were mixed with gDNA at contamination ratios of 1%, 2%, 3%, 4%, 5%, 10%, 15%, 20%, 25%, 30%, 35%, 40%, 45%, and 50% before library construction and sequencing.

[0076] The results of the cross-sample contamination assessment of the experimental simulation data are shown in Table 1 and Figures 4-18 As shown.

[0077] Table 1

[0078]

[0079]

[0080] Note: Sample: Simulated data with different contamination ratios; Total_num: SNP loci meeting the above requirements; Valid_num: SNP loci with an alt allele frequency between 0 and 0.005, theoretically normal loci; Valid_ratio: Valid_num / Total_num; AF_median: Median of SNP loci with AF values ​​greater than 0; 2*AF_median: Represents the contamination level of the sample to some extent; Contamination_ratio: Contamination level prediction result.

[0081] Table 1 shows the simulated data with different contamination ratios in the "Sample" column, and the "Contamination_ratio" represents the predicted contamination level. These samples were simulated with data at 1%, 2%, 3%, 4%, 5%, 10%, 15%, 20%, 25%, 30%, 35%, 40%, 45%, and 50% contamination ratios, respectively. The predicted contamination levels obtained by the model were 0, 0.84, 1.84, 2.72, 4.11, 4.67, 10.06, 16.08, 21.16, 26.68, 30.15, 36.55, 41.68, 46.07, and 47.86, respectively. The Pearson correlation coefficient between the predicted results and the simulated ratios reached 0.9983699, demonstrating that the model established by this method can accurately predict contamination levels and identify contaminated samples.

[0082] Figures 4-18 The horizontal axis represents the SNP sites used for pollution assessment, and the vertical axis represents the AF dose change at each SNP site. The positions of the four red dashed lines represent AF dose changes of 0.025, 0.05, 0.10, and 0.20, respectively.

[0083] Example 4

[0084] This embodiment uses the model constructed in Embodiment 2 to assess the pollution level of simulated data.

[0085] Whole-exome sequencing data of the samples were simulated with contamination rates of 1% (1M and 99M), 2% (2M and 98M), 3% (3M and 97M), 4% (4M and 96M), 5% (5M and 95M), 10% (10M and 90M), 15% (15M and 85M), 20% (20M and 80M), 25% (25M and 75M), 30% (30M and 70M), 35% (35M and 65M), 40% (40M and 60M), 45% (45M and 55M), and 50% (50M and 50M).

[0086] The results of the cross-sample contamination assessment of the simulated data are shown in Table 2.

[0087] Table 2

[0088]

[0089] Note: Sample: Simulated data with different contamination ratios; Total_num: SNP loci meeting the above requirements; Valid_num: SNP loci with an alt allele frequency between 0 and 0.005, theoretically normal loci; Valid_ratio: Valid_num / Total_num; AF_median: Median of SNP loci with AF values ​​greater than 0; 2*AF_median: Represents the contamination level of the sample to some extent; Contamination_ratio: Contamination level prediction result.

[0090] Table 2 shows the simulated data with different pollution ratios in the "Sample" column, and the "Contamination_ratio" represents the predicted pollution level. These samples were simulated with data at 1%, 2%, 3%, 4%, 5%, 10%, 15%, 20%, 25%, 30%, 35%, 40%, 45%, and 50% pollution levels, respectively. The predicted pollution levels obtained by the model were 0, 1.02, 2.01, 2.99, 3.99, 5.03, 10.65, 15.78, 21.09, 25.64, 29.38, 35.65, 40.88, 45.57, and 46.83. The Pearson correlation coefficient between the predicted results and the simulated ratios reached 0.9983729, demonstrating that the model established in this invention can accurately predict pollution levels and screen out contaminated samples.

[0091] In summary, this invention constructs a model by screening stable SNP loci simulation data with high population frequency, and assesses the degree of contamination based on the model. The AF values ​​in different samples are relatively stable and concentrated around 0, 0.5 and 1, which can accurately reflect the degree of cross-contamination of samples. The operation is simple, fast and convenient.

[0092] The applicant declares that the detailed method of the present invention is illustrated by the above embodiments, but the present invention is not limited to the above detailed method, that is, it does not mean that the present invention must rely on the above detailed method to be implemented. Those skilled in the art should understand that any improvements to the present invention, equivalent substitutions of the raw materials of the product of the present invention, addition of auxiliary components, selection of specific methods, etc., all fall within the protection scope and disclosure scope of the present invention.

Claims

1. A method for detecting cross-contamination of samples based on SNP sites, characterized in that, The method includes: Screening SNP sites, simulating data, building models, and assessing contamination levels based on the models; The screening of SNP sites includes the following steps: (1) Download the sequencing-related FASTQ file or bam file from the database; (2) Perform mutation detection on the high-throughput sequencing data of each sample in the file to obtain the mutation information of the sample in gvcf format file; (3) Merge all sample gvcf files to obtain the vcf format variation information of all samples; (4) Screen SNP sites based on the frequency of heterozygous populations and the frequency of variant alleles; The screening of SNP sites in step (4) includes: retaining the target variant, deleting the INDEL site, retaining the SNP sites whose FILTER column is PASS and which meet the requirements of heterozygous population frequency and untyped proportion, calculating the standard deviation of the AF value of each SNP site in the heterozygous state and the standard deviation of the AF value in the homozygous state in the normal population, summing them, and finally sorting them from smallest to largest according to the summed values, selecting sites with an AF value between 0.4-0.6 and 0.9-1 with a proportion greater than 0.30; The methods for constructing the model include univariate linear regression or Loess regression; The univariate linear regression includes: calculating the dose change of AF value to 0, 0.5 and 1 for each SNP site, and retaining the site closest to 0; The formula for the model is: Where y is the actual simulated pollution proportion, x is the degree of pollution, calculated using AF×2, α is the regression coefficient, and β is the intercept, calculated from the simulation data.

2. The method for detecting cross-contamination of samples based on SNP sites according to claim 1, characterized in that, The sequencing described in step (1) includes any one of whole genome sequencing, whole exome sequencing, or target sequence capture high-throughput sequencing.

3. The method for detecting cross-contamination of samples based on SNP sites according to claim 1, characterized in that, The target variants include conventional all external variants.

4. The method for detecting cross-contamination of samples based on SNP sites according to claim 1, characterized in that, The frequency of the heterozygous population is 0.3-0.

7.

5. The method for detecting sample cross-contamination based on SNP sites according to claim 1, characterized in that, The unclassified proportion is <0.10%.

6. The method for detecting cross-contamination of samples based on SNP sites according to claim 1, characterized in that, The number of sites is 200-2000.

7. The method for detecting cross-contamination of samples based on SNP sites according to claim 1, characterized in that, The frequency of the variant alleles mentioned in step (4) is 0.05-1.

8. The method for detecting cross-contamination of samples based on SNP sites according to claim 1, characterized in that, The simulated data includes: aligning sample mixed reads to a reference genome according to an algorithm, generating alignment files in BAM or SAM format, and then reordering and removing duplicate reads from the alignment files; the algorithm includes the BWA MEM algorithm.

9. The application of the method according to any one of claims 1-8 in detecting cross-contamination of samples.

Images