Tumor single sample somatic mutation discrimination and single sample tumor mutation load detection method based on high-throughput sequencing

Through high-throughput sequencing technology, normal population databases and tumor sample databases are constructed to screen somatic mutations, solving the problem of single-sample TMB detection, and achieving accurate tumor mutation burden assessment, supporting decision-making in immunotherapy.

CN118932056APending Publication Date: 2024-11-12NANODIGMBIO (NANJING) BIOTECHNOLOGY CO LTD

Patent Information

Application Number
CN202410989156.0
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-22
Publication Date
2024-11-12

AI Technical Summary

Technical Problem

In the prior art, tumor mutation load (TMB) detection requires the acquisition of tumor samples and normal control samples at the same time, and there is a problem that control samples are difficult to obtain or improperly preserved, resulting in the inability to conduct single-sample detection, which affects the judgment of immunotherapy.

Method used

Through a high-throughput sequencing method, a normal population database was constructed, exon capture sequencing and quality control of tumor samples was performed, and somatic mutations were screened, tumor single sample purity and mutation load were calculated, and a single sample purity and mutation load were constructed. A single-sample tumor mutation burden assessment model to realize single-sample TMB detection.

Benefits of technology

The problem of single-sample detection is overcome, the TMB value can be accurately calculated, and the evidence basis for immunotherapy is provided. The calculation results are highly correlated with paired sample detection, which is suitable for patients who cannot obtain control samples.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118932056A_ABST
    Figure CN118932056A_ABST
Patent Text Reader

Abstract

The invention belongs to the technical field of bioinformatics and high-throughput sequencing, and particularly relates to a tumor single-sample somatic mutation discrimination and single-sample tumor mutation load detection method based on high-throughput sequencing. The method comprises the following steps: constructing a necessary database through high-throughput sequencing data of a normal group sample, analyzing the high-throughput sequencing data of a single tumor sample in combination with a plurality of known public databases and common false positive site characteristics, distinguishing somatic mutation, and finally calculating a TMB numerical value. Therefore, the purposes of distinguishing the somatic mutation of the tumor single sample and detecting the TMB are achieved. The TMB value can also be calculated for single tumor data of a contrast sample which cannot be obtained, the technical problem that in the prior art, paired samples must be used for TMB value detection is solved, and a certain evidence basis is provided for judging whether a patient is suitable for immunotherapy or not.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the technical field of bioinformatics and high-throughput sequencing, and specifically relates to a method for distinguishing somatic mutations in a single tumor sample and detecting tumor mutation load in a single sample based on high-throughput sequencing. Background Art

[0002] Human individuals develop from fertilized eggs. Mutations in fertilized eggs are called germline mutations, which are mainly inherited from both parents. After an individual grows and develops, new mutations occur in his or her own cells due to stimulation from many external factors. These mutations are called somatic mutations. Somatic mutations can be divided into passenger mutations and driver mutations according to their harmfulness. Most passenger mutations are benign mutations that do not affect gene function, protein structure, etc. A small number of them will cause benign tumors through signal pathways, regulatory elements, etc., and then cause driver mutations under the influence of certain environmental factors, leading to the transformation of benign tumors into malignant tumors, namely cancer. Driver mutations are basically located in important genes such as individual growth, individual development, and DNA repair, and driver mutations will cause abnormalities in the growth, division, morphology, and function of normal cells, thereby driving the transformation of normal cells into cancer cells.

[0003] There are more than 1,000 types of cancer. In recent years, immunotherapy has shown good clinical effects in many cancers. Tumor mutation burden (TMB), as one of the markers of immunotherapy, is defined as the total number of somatic gene coding errors, base substitutions, gene insertions or deletions detected per million bases. The TMB value can be judged as TMB-H (TMB high) and TMB-L (TMB low) according to the threshold. Immunotherapy may be more effective for patients with TMB-H.

[0004] The current TMB detection method requires the same patient's tumor sample and normal control sample to be sequenced together, and then the mutations detected in the control sample are removed from the tumor sample to obtain somatic mutations, and then the TMB value is calculated. However, due to factors such as the difficulty in obtaining adjacent control samples for certain cancer patients, improper storage of control samples, patient reasons, and the method of obtaining control samples, the method that requires paired samples for TMB detection is difficult to perform. Summary of the invention

[0005] The purpose of the present invention is a method for distinguishing somatic mutations in a single tumor sample and detecting tumor mutation load in a single sample based on high-throughput sequencing. The method can detect the tumor mutation load of a single tumor sample.

[0006] The present invention provides a method for distinguishing somatic mutations in a single tumor sample based on high-throughput sequencing, comprising the following steps:

[0007] Constructing a normal population database based on high-throughput sequencing data of normal population samples, wherein the normal population database includes a normal population SNP information database, a normal population germline mutation database, and deep data within the normal population capture region;

[0008] Perform exon capture sequencing and quality control on tumor samples, compare and filter the quality-controlled tumor sample sequencing data with the human reference genome to obtain the tumor sample bam file;

[0009] Perform SNP analysis and deep capture analysis based on the tumor sample bam file to obtain SNP sites of the tumor sample and deep capture area data of the tumor sample, respectively;

[0010] Calculate the purity of a single tumor sample based on the tumor sample SNP site, the tumor sample capture region depth data, the normal population SNP information database and the normal population capture region depth data;

[0011] Perform mutation detection based on the tumor sample bam file to obtain a tumor sample mutation vcf file;

[0012] The tumor sample mutation vcf file is screened by normal population germline mutation database, human public database and false positive mutation feature to obtain somatic mutation results;

[0013] The somatic mutation results are filtered based on the purity of the single tumor sample, the population frequency and the tumor sample capture area depth data to obtain the somatic mutation data.

[0014] Preferably, the construction of the normal population database comprises: performing exon capture sequencing and quality control on normal population samples, comparing and filtering the sequencing data of the normal population samples after quality control with the human reference genome, and obtaining a bam file of the normal population samples;

[0015] Perform SNP analysis and deep capture analysis based on the normal population sample bam file to obtain a normal population sample SNP information database and deep data within the normal population capture area, respectively;

[0016] Based on the bam file of the normal population samples, mutation detection is performed to obtain a normal population germline mutation database.

[0017] Preferably, the screening of the normal population germline mutation database includes: based on the tumor sample mutation vcf file, filtering and removing the germline mutations in the normal population germline mutation database to obtain the normal population germline mutation data with the germline mutations removed.

[0018] Preferably, the human public database screening includes: based on the normal population germline mutation data from the removal of germline mutations, removing mutation sites with a population frequency greater than 1% in the exome sequencing data of the gnomAD database, removing mutation sites marked as "COMMON" in the dbSNP database, and removing mutation sites marked as "germline" origin in the Clinvar database.

[0019] Preferably, after screening the human public database, the false positive mutation signature screening includes at least one of the following:

[0020] Remove mutations located in the human leukocyte antigen gene;

[0021] Remove multiple different genotype mutations at the same genomic location;

[0022] Remove mutations with read support number less than 5 and total site depth less than 20;

[0023] Remove mutations located in single-base and three-base repeat regions with a mutation frequency lower than 5%, remove mutations located in two-base repeat regions with a mutation frequency lower than 4%, and remove mutations located in four-base repeat regions with a mutation frequency lower than 3%;

[0024] Remove mutations where the mutation site is located 10 bp to the left and right of the softclip region and the reads account for more than 20% of the total mutation reads;

[0025] Remove mutations with low alignment quality less than 10 and mutations that account for more than 50% of the total mutation reads;

[0026] Remove mutations whose mutation sites are located 10 bp to the left and right of the read edge and whose reads account for more than 50% of the total mutation reads;

[0027] Remove non-hotspot mutations with oxidative damage and DNA damage and mutation frequency less than 3% and hotspot mutations with mutation frequency less than 2%;

[0028] Remove mutations with a positive-strand to negative-strand read ratio greater than 9 in the mutation reads.

[0029] Preferably, obtaining the purity of the single tumor sample comprises the following steps:

[0030] In the normal population SNP information database, the reads support number of each detected site ref and alt is counted and summed, and the total read number after the sum is divided by the total number of normal population samples to obtain the overall SNP information of the normal population;

[0031] Counting the depth mean of the normal population samples within the 100 bp sliding window of the normal population capture region to obtain the depth mean within the normal population capture region;

[0032] Supplementing the SNP site in the overall SNP information of the normal population that has the same position as the SNP site of the tumor sample to the SNP site of the tumor sample to obtain the supplemented SNP site of the tumor sample;

[0033] Supplementing the overall depth mean of the normal population at the same position as the 100 bp sliding window of the tumor sample capture region into the 100 bp sliding window of the tumor sample capture region to obtain supplemented tumor sample depth information;

[0034] The scarHRD software was used to analyze the supplemented tumor sample SNP sites and the supplemented tumor sample depth information, and the purity of the tumor sample was obtained according to the copy number variation state distribution corresponding to the mutation site.

[0035] Preferably, the normal population samples include one or more of normal human tissue samples, normal human blood cell samples, tumor patient blood cell samples and tumor patient adjacent cancer samples.

[0036] Preferably, the version of the human reference genome includes the hg19 genome.

[0037] The present invention also provides a method for constructing a single-sample tumor mutation load assessment model, based on the discrimination method described in the above technical solution to obtain somatic mutation data and exon capture sequencing capture regions, and calculate the tumor mutation load according to a predetermined formula (1):

[0038] TMB = mut / panel size, formula (1);

[0039] Wherein, mut refers to the number of somatic mutations included in the somatic mutation data, and panel size refers to the size of the captured region of exon capture sequencing, in Mb.

[0040] The present invention also provides a computer-readable storage medium, on which a computer program or instruction is stored, and wherein the computer program or instruction, when executed by a processor, implements the steps of the method described in the above technical solution.

[0041] The present invention also provides a computer program product, including a computer program or instructions, characterized in that when the computer program or instructions are executed by a processor, the steps of the method described in the above technical solution are implemented.

[0042] The present invention also provides the use of the method described in the above technical solution, the computer-readable storage medium described in the above technical solution, or the computer program product described in the above technical solution in the preparation of a product for guiding tumor immunotherapy.

[0043] Beneficial effects:

[0044] The present invention provides a method for distinguishing somatic mutations in a single tumor sample based on high-throughput sequencing. On this basis, the present invention also provides a method for constructing a single sample tumor mutation burden (TMB) evaluation model. The present invention first constructs a necessary database through high-throughput sequencing data of normal population samples, combines several known public databases and common false positive site characteristics, analyzes the high-throughput sequencing data of a single tumor sample, distinguishes somatic mutations, and finally calculates the TMB value to achieve the purpose of distinguishing somatic mutations in a single tumor sample and detecting TMB. The TMB value can also be calculated for single tumor data that cannot obtain a control sample, which overcomes the technical problem that paired samples must be used to detect TMB values ​​in the prior art, and provides a certain evidence basis for determining whether a patient is suitable for immunotherapy. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required to be used in the embodiments are briefly introduced below.

[0046] Figure 1 This is a flow chart of the method for identifying somatic mutations in a single human tumor sample and detecting TMB in a single sample based on next-generation sequencing in Example 3;

[0047] Figure 2 A correlation diagram of TMB values ​​calculated for paired TMB standards and single samples and reference TMB values ​​in Example 2;

[0048] Figure 3 This is the purity correlation diagram of single and double samples of real samples in Example 3.

[0049] Figure 4 This is the TMB correlation diagram of the real sample single and double samples in Example 3. DETAILED DESCRIPTION

[0050] The present invention provides a method for distinguishing somatic mutations in a single tumor sample based on high-throughput sequencing, comprising the following steps:

[0051] Constructing a normal population database based on high-throughput sequencing data of normal population samples, wherein the normal population database includes a normal population SNP information database, a normal population germline mutation database, and deep data within the normal population capture region;

[0052] Perform exon capture sequencing and quality control on tumor samples, compare and filter the quality-controlled tumor sample sequencing data with the human reference genome to obtain the tumor sample bam file;

[0053] Perform SNP analysis and deep capture analysis based on the tumor sample bam file to obtain SNP sites of the tumor sample and deep capture area data of the tumor sample, respectively;

[0054] Calculate the purity of a single tumor sample based on the tumor sample SNP site, the tumor sample capture region depth data, the normal population SNP information database and the normal population capture region depth data;

[0055] Perform mutation detection based on the tumor sample bam file to obtain a tumor sample mutation vcf file;

[0056] The tumor sample mutation vcf file is screened by normal population germline mutation database, human public database and false positive mutation feature to obtain somatic mutation results;

[0057] The somatic mutation results are filtered based on the purity of the single tumor sample, the population frequency and the tumor sample capture area depth data to obtain the somatic mutation data.

[0058] The present invention constructs a normal population database based on high-throughput sequencing data of normal population samples, and the normal population database includes a normal population SNP information database, a normal population germline mutation database, and a normal population capture region depth data. The normal population samples of the present invention preferably include but are not limited to one or more of normal human tissue samples, normal human blood cell samples, tumor patient blood cell samples, and tumor patient adjacent cancer samples; the blood cell samples preferably include plasma cell samples, more preferably plasma leukocyte samples; the number of normal population samples is not specifically limited, but the more samples, the more accurate the test results, such as the normal population samples used in one embodiment of the present invention are 57 cases. In the process of constructing the normal population database, the high-throughput sequencing preferably includes second-generation sequencing, more preferably exon capture sequencing based on second-generation sequencing; the depth of the high-throughput sequencing is preferably>100X; the specific method of the high-throughput sequencing is not specifically limited, and the conventional high-throughput sequencing method in the art can be used.

[0059] On this basis, the method for constructing the normal population database specifically preferably includes: performing exon capture sequencing and quality control on normal population samples, and comparing and filtering the normal population sample sequencing data after quality control with the human reference genome to obtain a normal population sample bam file; performing SNP analysis and deep capture analysis based on the normal population sample bam file to obtain a normal population sample SNP information database and a normal population sample capture region data, respectively; performing variation detection based on the normal population sample bam file to obtain a normal population germline mutation database.

[0060] The present invention preferably performs exon capture sequencing and quality control on normal population samples, and compares and filters the sequencing data of normal population samples after quality control with the human reference genome to obtain a bam file of normal population samples. The software used for quality control in the present invention preferably includes fastp software, and the quality control can remove low-quality data; the software used for the comparison preferably includes bwa software. The human reference genome in the present invention preferably includes the hg19 genome; the software used for the filtering preferably includes gatk software, and the gatk software can compare data markers for duplication and base quality correction to obtain a bam file of normal population samples. The present invention does not specifically limit the specific operation process of the fastp software, bwa software and gatk software, and can be performed according to the conventional operation process of those skilled in the art.

[0061] After obtaining the normal population sample bam file, the present invention preferably performs SNP analysis and deep capture analysis based on the normal population sample bam file to obtain a normal population sample SNP information database and normal population sample capture area data, respectively. The present invention preferably uses varscan software to perform the SNP analysis, and uses samtools software to perform the deep capture analysis to obtain the depth mean of the normal population sample in the capture area.

[0062] At the same time, the present invention preferably performs variation detection based on the bam file of the normal population sample to obtain a normal population germline mutation database. The variation detection of the present invention preferably uses the GenomicsDBImport and CreateSomaticPanelOfNormals modules in the gatk software to construct a normal population germline mutation vcf file, that is, the normal population germline mutation database.

[0063] The present invention performs exon capture sequencing and quality control on tumor samples, and compares and filters the quality-controlled tumor sample sequencing data with the human reference genome to obtain a tumor sample bam file. The process of performing exon capture sequencing, quality control, comparison and filtering on the tumor sample in the present invention is preferably the same as the parameters in the process of obtaining the bam file of the normal population sample described above, and will not be repeated.

[0064] After obtaining the tumor sample bam file, the present invention performs SNP analysis and deep capture analysis based on the tumor sample bam file to obtain the tumor sample SNP site and tumor sample capture area depth data respectively. The process of SNP analysis and deep capture analysis of the tumor sample of the present invention is preferably the same as that of the normal population sample, and will not be repeated.

[0065] After obtaining the tumor sample SNP site and tumor sample capture area depth data, the present invention calculates the purity of the tumor single sample based on the tumor sample SNP site, tumor sample capture area depth data, normal population SNP information database and normal population capture area depth data.

[0066] The method for calculating the purity of a single tumor sample described in the present invention preferably includes the following steps: in the normal population SNP information database, counting and summing the reads support number of each detected site ref and alt, dividing the total number of reads after the sum by the total number of normal population samples, to obtain the normal population overall SNP information; counting the depth mean of the normal population samples within the 100bp sliding window of the normal population capture region, to obtain the depth mean within the normal population capture region; supplementing the SNP site in the normal population overall SNP information that has the same position as the tumor sample SNP site to the tumor sample SNP site, to obtain the supplemented tumor sample SNP site; supplementing the normal population overall depth mean at the same position as the 100bp sliding window of the tumor sample capture region to the 100bp sliding window of the tumor sample capture region, to obtain the supplemented tumor sample depth information; using scarHRD software for analysis based on the supplemented tumor sample SNP site and the supplemented tumor sample depth information, and obtaining the purity of the tumor sample based on the copy number variation state distribution corresponding to the mutation site.

[0067] At the same time, the present invention performs mutation detection based on the tumor sample bam file to obtain the tumor sample mutation vcf file. The mutation detection of the present invention preferably includes using the mutect2 module of the gatk software to detect mutations to obtain the tumor sample mutation vcf file.

[0068] After obtaining the tumor sample mutation vcf file, the present invention performs normal population germline mutation database screening, human public database screening and false positive mutation feature screening on the tumor sample mutation vcf file to obtain a somatic mutation file. The normal population germline mutation database screening, human public database screening and false positive mutation feature screening are not limited in time sequence. That is, in one embodiment of the present invention, the screening process is described by taking the order of performing normal population germline mutation database screening, human public database screening and false positive mutation feature screening on the tumor sample mutation vcf file in sequence as an example, but this screening order cannot be regarded as the entire protection scope of the screening described in the present invention.

[0069] The screening of the normal population germline mutation database of the present invention preferably includes: filtering and removing the germline mutations in the normal population germline mutation database.

[0070] Based on the data after screening the normal population germline mutation database, the present invention performs human public database screening on the screened data, and the method for screening the human public database preferably includes: removing mutation sites with a population frequency greater than 1% in the exon sequencing data of the gnomAD database, removing mutation sites with the standard of "COMMON" in the dbSNP database, and removing mutation sites marked as "germline" sources in the Clinvar database. The gnomAD database of the present invention is preferably the gnomAD v2.1.1 version; the dbSNP database is preferably the dbSNP155 version database. The present invention does not specifically limit the specific process of screening the human public database, and conventional operations can be performed according to the above-defined step parameters. The steps of screening the human public database of the present invention can remove as many germline mutations as possible in single tumor samples.

[0071] Based on the data screened from the human public database, the present invention performs false positive mutation feature screening on the screened data, and the false positive mutation feature screening includes at least one of the following multiple items, and more preferably includes the following multiple items: removing mutations located in the human leukocyte antigen gene; removing multiple different genotype mutations at the same genomic position; removing mutations with a reads support number of less than 5 and a total site depth of less than 20; removing mutations located in single-base and three-base repeat regions with a mutation frequency of less than 5%, removing mutations located in two-base repeat regions with a mutation frequency of less than 4%, and removing mutations located in four-base repeat regions with a mutation frequency of less than 3%; removing mutations where the mutation site is located 10 bp to the left and right and the reads in the softclip region account for more than 20% of the total mutation reads; removing mutations where the mutation reads with low alignment quality less than 10 account for more than 50% of the total mutation reads; removing mutations where the mutation site is located 10 bp to the left and right of the reads edge and the reads account for more than 10% of the total mutation reads. More than 50% mutations; remove non-hotspot mutations with oxidative damage and DNA damage and mutation frequency less than 3% and hotspot mutations with mutation frequency less than 2%; remove mutations with a positive-strand to negative-strand read ratio greater than 9 in mutation reads. It is well known to those skilled in the art that many false mutations will be introduced during library construction and sequencing, such as low-frequency mutations introduced by enzyme digestion library construction, false mutations introduced by PCR amplification during library construction, and false mutations in repeated regions during library construction and sequencing. The false positive mutation feature screening step of the present invention can remove false mutations, so that as many true mutations from tumor cells as possible are included in the construction of the tumor single sample somatic cell mutation discrimination model.

[0072] After obtaining the somatic mutation file, the present invention filters the somatic mutation file based on the purity of the single tumor sample, the population frequency and the tumor sample capture area depth data, and constructs a tumor single sample somatic mutation discrimination model with the filtered somatic mutation data to obtain a somatic mutation discrimination result.

[0073] The present invention also provides a method for constructing a single-sample tumor mutation load assessment model, based on the discrimination method described in the above technical solution, somatic mutation data and exon capture sequencing capture regions are obtained and tumor mutation load is calculated according to a predetermined formula (1):

[0074] TMB=mut / panel size, formula (1); wherein mut refers to the number of somatic mutations included in the somatic mutation data, and panel size refers to the size of the exon capture sequencing capture region, in Mb.

[0075] The present invention also provides a computer-readable storage medium, on which a computer program or instruction is stored. When the computer program or instruction is executed by a processor, the steps of the method described in the above technical solution are implemented.

[0076] The present invention also provides a computer program product, including a computer program or instructions, which implement the steps of the method described in the above technical solution when executed by a processor.

[0077] The present invention also provides the use of the method described in the above technical solution, the computer-readable storage medium described in the above technical solution, or the computer program product described in the above technical solution in the preparation of a product for guiding tumor immunotherapy. Based on the above-mentioned tumor single sample somatic mutation discrimination and single sample tumor mutation load detection method based on high-throughput sequencing, the TMB value detection of a single tumor sample can be realized, and it has been proved through experiments that the calculated TMB value has a high correlation with the TMB value obtained by paired sample detection and the TMB value of the standard product, which can reach 0.8885 or more, indicating that the method for calculating the TMB value of a single tumor sample described in the present invention is accurate and practical, and can be used for immunotherapy and guidance of drugs during immunotherapy.

[0078] In order to further illustrate the present invention, the technical solution provided by the present invention is described in detail below in conjunction with the accompanying drawings and embodiments, but they should not be construed as limiting the protection scope of the present invention.

[0079] Example 1

[0080] The steps of human tumor single sample somatic mutation discrimination and single sample TMB detection method based on second-generation sequencing are as follows:

[0081] 1. Tumor sample DNA is library-built using whole exome capture reagents and subjected to next-generation sequencing;

[0082] 2. Use fastp software to perform quality control on the second-generation sequencing data, remove low-quality data, and then use bwa software to align the data to the human reference genome hg19. Use gatk software to compare the data for duplicate marking and base quality correction to obtain the bam file for subsequent analysis;

[0083] 3. The normal population samples were subjected to the same steps 1 and 2 above to obtain the analyzed bam files. Mutect2 was used to detect mutations in the bam files of the normal population samples. The GenomicsDBInport and CreateSomaticPanelOfNormals modules were used to construct the normal population PON mutation files for the population mutation results.

[0084] 4. Use mutec2 to detect mutations in single tumor samples and filter out germline mutations detected in the PON file of the normal population;

[0085] 5. For the mutations filtered in 4 above, use the known common database to filter again. The filtering conditions are as follows:

[0086] 5a. Remove mutation sites with a population frequency greater than 1% in the Exomes data of gnomAD v2.1.1;

[0087] 5b. Remove the mutation sites marked as "COMMON" in the dbSNP155 version database;

[0088] 5c. Remove mutation sites annotated as “germline” in the Clinvar database.

[0089] 6. For the mutations filtered in step 5 above, use the false positive mutation features to filter again. The filtering conditions are as follows:

[0090] 6a. Remove mutations in the human leukocyte antigen gene;

[0091] 6b. Remove multiple different genotype mutations at the same genomic location;

[0092] 6c. Remove mutations with read support less than 5 and total site depth less than 20;

[0093] 6d. Remove mutations located in single-base and three-base repeat regions with a mutation frequency below 5%, remove mutations located in two-base repeat regions with a mutation frequency below 4%, and remove mutations located in four-base repeat regions with a mutation frequency below 3%;

[0094] 6e. Remove mutations where the reads located in the softclip region 10 bp to the left and right of the mutation site account for more than 20% of the total mutation reads;

[0095] 6f. Remove mutations with low alignment quality less than 10 and whose mutation reads account for more than 50% of the total mutation reads;

[0096] 6g. Remove mutations whose mutation sites are located 10 bp to the left and right of the read edge and account for more than 50% of the total mutation reads;

[0097] 6h. Remove oxidative damage (DD) OD and DNA damage (OD) with mutation frequency less than 3% non-hotspot mutations and hotspot mutations with mutation frequency less than 2%;

[0098] 6i. Remove mutations with a positive-strand to negative-strand read ratio greater than 9 in mutation reads;

[0099] 7. Calculate the purity of the tumor sample based on the SNP and depth information of the normal population sample. The purity calculation method is as follows:

[0100] 7a. Use varscan software to detect SNPs in pairs of normal population samples, and count the number of reads supporting each ref and alt site detected in the normal population. The sum is divided by the total number of normal samples to obtain the overall SNP information of the normal population;

[0101] 7b.samtools overall depth mean of the control group within the 100bp sliding window of the whole exon capture region for normal samples;

[0102] 7c. The corresponding SNP information of the normal population is supplemented according to the SNP sites of the tumor sample, and the area within the 100bp sliding window of the whole exon capture region of the tumor sample is supplemented with the deep information of the normal population;

[0103] 7d. Based on the tumor SNP site information supplemented with the control population information and the 100bp sliding window region depth information of the whole exon capture region, the scarHRD software was used to analyze the purity of the tumor sample;

[0104] 8. For the mutations filtered in step 6 above, filter the germline mutations in the tumor sample according to the tumor purity in step 7 above;

[0105] 9. Filter the mutations filtered in step 8 above according to the capture area panel and population frequency;

[0106] 10. Calculate TMB for the mutations filtered in step 9 above. The calculation method is as follows:

[0107] TMB = mut / (panel size (Mb))

[0108] Where mut refers to the number of somatic mutations included in the calculation of TMB, and panel size (Mb) refers to the coding size of the captured panel 1 in the exon region, in Mb.

[0109] Example 2

[0110] In this example, 6 tumor gDNA TMB standards and 6 control sample gDNA were captured and sequenced using a whole exome panel, and the sequencing depth was about 600X. The standards were purchased from Jingliang Technology Co., Ltd. and the TMB reference values ​​are shown in the following table.

[0111] Table 16 TMB reference values ​​for TMB standards

[0112] sample Standards TMB06T 23.26 TMB08T 16.62 TMB14T 6.21 TMB16T 9.03 TMB18T 3.18 TMB20T 4.69

[0113] 1.6 TMB standard tumor and control sample sequencing data were removed by fastp (version 0.20.0) software with sequencing quality less than Q15, base quality unqualified greater than 40%, N bases more than 5, and sequencing sequences less than 15bp in length after trimming low-quality bases. Then, bwa (version 0.7.17) software was used to align the human UCSC hg19 genome, and gatk (version 4.2.6.1) software was used to mark the duplicate sequences in the alignment file and re-calibrate the base quality. The obtained bam format file was used as the input file for single-sample somatic mutation discrimination and single-sample TMB calculation analysis.

[0114] 2.6 TMB standards were calculated for purity with paired samples and single sample purity. The paired sample purity was analyzed using VarScan (version 2.4.3) software for SNP and copy number variation regions, and scarHRD (version 0) was used to analyze SNP and copy number variation regions to obtain sample purity. The calculation method of sample purity was the same as in Example 1, except that the information of the normal population was provided by paired samples;

[0115] The single sample purity is calculated using the calculation method of step 7 in Example 1.

[0116] The consistency of purity calculated by single samples and paired samples is shown in Table 2. The results show that the purity calculated by single samples of 6 TMB standards is highly consistent with the purity calculated by paired samples. This result indicates that the purity calculation results of single samples can be used for filtering analysis of somatic cells.

[0117] Table 26 Paired and Single Sample Purity Calculations of TMB Standards

[0118] sample Purity of pairwise calculations Purity calculated from a single sample TMB06T 1.00 0.91 TMB08T 0.98 0.96 TMB14T 0.99 0.97 TMB16T 0.99 0.93 TMB18T 0.99 0.96 TMB20T 0.99 0.99

[0119] 3.6 TMB standards were used for somatic mutation detection with paired samples and calculation of TMB values, and single sample somatic mutation detection and calculation of TMB values.

[0120] 3.1 The bam files of the paired samples were analyzed for germline mutations using the HaplotypeCaller module in the gatk software, and the mutect2 module in the gatk software was used to analyze somatic mutations. The preliminary results of somatic mutations obtained were removed from germline mutations, PON mutations, mutations at sites with a gnomad frequency greater than 0.01, dbSNP155 common mutations, and false positive mutation characteristic mutations, and the TMB value was calculated according to the formula TMB=mut / (panel size (Mb)).

[0121] 3.2 The TMB value of a single sample was calculated using the method and formula in Example 1.

[0122] 4. The TMB values ​​calculated using paired samples and single samples were analyzed for correlation with the TMB reference values ​​of the 6 tumor samples. The results are as follows: Figure 2 As shown, in Figure 2 The horizontal axis represents the TMB reference value, and the vertical axis represents the TMB value calculated from the paired samples of high-throughput sequencing data.

[0123] Depend on Figure 2 It can be concluded that the TMB values ​​calculated by single samples and paired samples are consistent. The results show that the TMB values ​​calculated by paired samples of 6 TMB standard products are correlated with the TMB reference values. 2 The value can reach 0.9848, indicating that the method of calculating the TMB value of paired samples of NGS data can reflect the tumor mutation load status of the sample, and the correlation R between the TMB value calculated by a single sample and the TMB reference value calculated is 2 It is 0.9166, indicating that the TMB value calculated by a single sample is strongly correlated with the TMB reference value.

[0124] Example 3

[0125] The steps of the single-sample somatic mutation discrimination and single-sample TMB detection method for ovarian tumors based on second-generation sequencing are as follows: The technical roadmap is as follows: Figure 1 shown.

[0126] In this example, 57 real ovarian cancer samples and 57 control samples, the tumor samples are FFPE samples, and the control samples are blood gDNA, all of which are captured and sequenced using a whole exome panel, and the sequencing depth is about 150X. The 57 tumor and control samples were used to remove low-quality data using fastp software, the hg19 genome was aligned using bwa software, and the gatk software was used to perform repeat labeling and base quality recalibration to obtain the bam files for analysis. The purity of the bam files of the 57 tumor and control samples was calculated separately, and the purity of the paired samples and the single samples was calculated using the same method as in Example 2. The consistency of the purity of the paired samples (the 57 control samples were used as paired samples to calculate the purity) and the purity calculated for the single samples (the 57 control samples were used as normal population samples to calculate the purity) is as follows: Figure 3 As shown, the results show that the correlation R 2 The value is 0.8858, which indicates that the purity calculation result of a single sample can be used to filter out somatic cells.

[0127] As can be seen from Example 1, the TMB calculated by paired samples can reflect the true tumor mutation load status of the sample. For 57 tumor samples and control samples, paired analysis somatic cell TMB calculation (the same method as the paired TMB value calculation method in Example 2) and single tumor sample somatic cell TMB calculation were performed, and the results are as follows: Figure 4 As shown in (a), the results show that the correlation R 2 The value is 0.8885, indicating a strong correlation. Considering that the tumor sample is an ovarian cancer sample, a certain amount of normal samples may be contained in the tumor sample taken from the cancerous site. After removing the tumor sample with a tumor content of less than 30%, the correlation between the paired analysis of somatic cell TMB and the single tumor sample analysis of somatic cell TMB can reach 0.9320. The results are as follows Figure 4 As shown in (b), this indicates that this detection method can be used to perform TMB detection on a single tumor sample without a control sample. The higher the tumor content in the tumor sample, the more consistent the result is with the actual tumor mutation load status, which provides a certain evidence basis for immunotherapy for patients who have difficulty obtaining control samples.

[0128] Although the above embodiment describes the present invention in detail, it is only a part of the embodiments of the present invention, not all of the embodiments. People can also obtain other embodiments based on this embodiment without creativity, and these embodiments all fall within the protection scope of the present invention.

Claims

1. A method for identifying somatic mutations in a single tumor sample based on high-throughput sequencing, characterized in that: The steps include: Constructing a normal population database based on high-throughput sequencing data of normal population samples, wherein the normal population database includes a normal population SNP information database, a normal population germline mutation database, and deep data within the normal population capture region; Perform exon capture sequencing and quality control on tumor samples, compare and filter the quality-controlled tumor sample sequencing data with the human reference genome to obtain the tumor sample bam file; Perform SNP analysis and deep capture analysis based on the tumor sample bam file to obtain SNP sites of the tumor sample and deep capture area data of the tumor sample, respectively; Calculate the purity of a single tumor sample based on the tumor sample SNP site, the tumor sample capture region depth data, the normal population SNP information database and the normal population capture region depth data; Perform mutation detection based on the tumor sample bam file to obtain a tumor sample mutation vcf file; The tumor sample mutation vcf file is screened by normal population germline mutation database, human public database and false positive mutation feature to obtain somatic mutation results; The somatic mutation results are filtered based on the purity of the single tumor sample, the population frequency and the tumor sample capture area depth data to obtain the somatic mutation data.

2. The identification method according to claim 1, characterized in that: The construction of the normal population database comprises: performing exon capture sequencing and quality control on normal population samples, comparing and filtering the sequencing data of the normal population samples after quality control with the human reference genome, and obtaining a bam file of the normal population samples; Perform SNP analysis and deep capture analysis based on the normal population sample bam file to obtain a normal population sample SNP information database and deep data within the normal population capture area, respectively; Based on the bam file of the normal population samples, mutation detection is performed to obtain a normal population germline mutation database.

3. The identification method according to claim 1 or 2, characterized in that: The screening of the normal population germline mutation database includes: based on the tumor sample mutation vcf file, filtering and removing the germline mutations in the normal population germline mutation database to obtain the normal population germline mutation data without germline mutations.

4. The identification method according to claim 3, characterized in that: The human public database screening includes: based on the normal population germline mutation data from the removal of germline mutations, removing mutation sites with a population frequency greater than 1% in the exon sequencing data of the gnomAD database, removing mutation sites marked as "COMMON" in the dbSNP database, and removing mutation sites marked as "germline" in the Clinvar database.

5. The identification method according to claim 4, characterized in that: After screening the human public database, the false positive mutation signature screening includes at least one of the following: Remove mutations located in the human leukocyte antigen gene; Remove multiple different genotype mutations at the same genomic location; Remove mutations with read support number less than 5 and total site depth less than 20; Remove mutations located in single-base and three-base repeat regions with a mutation frequency lower than 5%, remove mutations located in two-base repeat regions with a mutation frequency lower than 4%, and remove mutations located in four-base repeat regions with a mutation frequency lower than 3%; Remove mutations where the mutation site is located 10 bp to the left and right of the softclip region and the reads account for more than 20% of the total mutation reads; Remove mutations with low alignment quality less than 10 and mutations that account for more than 50% of the total mutation reads; Remove mutations whose mutation sites are located 10 bp to the left and right of the read edge and whose reads account for more than 50% of the total mutation reads; Remove non-hotspot mutations with oxidative damage and DNA damage and mutation frequency less than 3% and hotspot mutations with mutation frequency less than 2%; Remove mutations with a positive-strand to negative-strand read ratio greater than 9 in the mutation reads.

6. The identification method according to claim 1, characterized in that: The step of obtaining the purity of the single tumor sample comprises the following steps: In the normal population SNP information database, the reads support number of each detected site ref and alt is counted and summed, and the total read number after the sum is divided by the total number of normal population samples to obtain the overall SNP information of the normal population; Counting the depth mean of the normal population samples within the 100 bp sliding window of the normal population capture region to obtain the depth mean within the normal population capture region; Supplementing the SNP site in the overall SNP information of the normal population that has the same position as the SNP site of the tumor sample to the SNP site of the tumor sample to obtain the supplemented SNP site of the tumor sample; Supplementing the overall depth mean of the normal population at the same position as the 100 bp sliding window of the tumor sample capture region into the 100 bp sliding window of the tumor sample capture region to obtain supplemented tumor sample depth information; The scarHRD software is used based on the supplemented tumor sample SNP sites and the supplemented tumor sample depth information, and the purity of the tumor sample is analyzed based on the copy number variation state distribution corresponding to the mutation site.

7. The identification method according to claim 1 or 2, characterized in that: The normal population samples include one or more of normal human tissue samples, normal human blood cell samples, tumor patient blood cell samples and tumor patient adjacent cancer samples.

8. The identification method according to claim 1, characterized in that: Versions of the human reference genome include the hg19 genome.

9. A method for constructing a single sample tumor mutation burden assessment model, characterized in that: Based on the discrimination method according to any one of claims 1 to 8, somatic mutation data and exon capture sequencing capture regions are obtained, and the tumor mutation load is calculated according to the predetermined formula (1): TMB = mut / panel size, formula (1); Wherein, mut refers to the number of somatic mutations included in the somatic mutation data, and panel size refers to the size of the captured region of exon capture sequencing, in Mb.

10. A computer-readable storage medium having a computer program or instruction stored thereon, characterized in that: When the computer program or instruction is executed by a processor, the steps of the method according to any one of claims 1 to 9 are implemented.

11. A computer program product comprising a computer program or instructions, characterized in that When the computer program or instruction is executed by a processor, the steps of the method according to any one of claims 1 to 9 are implemented.

12. Use of the method according to any one of claims 1 to 9, the computer-readable storage medium according to claim 10, or the computer program product according to claim 11 in the preparation of a product for guiding tumor immunotherapy.

Citation Information

Patent Citations

  • Tumor mutation analysis method and system, terminal and readable storage medium

    CN110570904A

  • Somatic mutation detection method and device for tumor tissue single samples

    CN111718982A

  • Method and device for evaluating homologous recombination defects of single sample and storage medium

    CN114067908A

Cited By

  • Leukemia gene mutation detection method based on next-generation sequencing transcriptome data

    CN119541636A