Detection method, electronic device and storage medium for micro-residual lesions based on blood ctDNA

Abstract

The present invention discloses a method, electronic device and storage medium for detecting micro-residual lesions based on blood ctDNA, and the detection method comprises the following steps: S1, panel sequencing data: cfDNA of a target region is amplified by primer combination, and sequencing is performed to obtain personalized panel sequencing data of a tumor patient; S2, the sequencing data is automatically analyzed; S3, MRD positive judgment: when the average mutation rate of the MRD monitoring site is significantly higher than the background mutation rate, the MRD result is considered to be positive, otherwise it is negative; S4, storing patient information and MRD results: using mysql database management system to store the basic information of the patient, as well as the monitoring time and results of this MRD, to form a continuous monitoring record of the patient, to form tracking of the patient and prognosis judgment. The detection method of the present invention can realize the monitoring of the MRD status of tumor patients at a relatively low cost, and the sensitivity of ctDNA detection reaches one ten-thousandth level.

Description

Technical Field

[0001] The present invention belongs to the technical field of gene detection, and in particular relates to a method, electronic device and storage medium for detecting tiny residual lesions based on blood ctDNA. Background Art

[0002] Minimal residual disease (MRD) refers to the state in which a small amount of tumor cells remain in the body of a cancer patient after treatment. These residual tumor cells may be very few in number, but they still exist and may cause tumor recurrence. MRD detection methods mainly include imaging detection based on CT or PET-CT, immunological detection based on tumor cell surface markers, and molecular biological detection based on gene mutations.

[0003] Circulating Tumor DNA (ctDNA) refers to the somatic DNA of tumor cells that is released into the circulatory system after shedding or apoptosis. It is a characteristic tumor biomarker. Through ctDNA testing, tumor traces in the blood can be detected. The applications of ctDNA mainly include:

[0004] (1) Tumor specificity: ctDNA carries tumor-specific mutation information and can reflect the genomic changes of the tumor.

[0005] (2) Dynamic monitoring: By detecting the level of ctDNA, the development and changes of the tumor can be dynamically monitored, which helps to evaluate the treatment effect and predict the risk of recurrence.

[0006] (3) Early diagnosis: ctDNA can be detected in the early stages of tumors, which helps with early diagnosis and early treatment.

[0007] (4) Prognosis evaluation: By detecting the type and number of ctDNA mutations, the patient's prognosis and survival can be evaluated.

[0008] Therefore, ctDNA, as an important tumor biomarker, can be used to evaluate the MRD status of cancer patients, comprehensively and accurately reflect the tumor burden level of cancer patients, help guide treatment plans, and thus improve patient prognosis.

[0009] However, ctDNA detection is limited by two major factors:

[0010] (1) The initial amount of ctDNA in plasma is very low;

[0011] (2) Current high-throughput sequencing technology has a high background error rate, which makes it difficult for conventional ctDNA detection to achieve a sufficiently low limit of detection (LOD), thereby affecting the accuracy of MRD assessment.

[0012] In summary, after cancer patients have received treatment and there are very few tumor cells in their bodies, using ctDNA to evaluate the MRD status of cancer patients requires a set of accurate and scientific experimental analysis methods to improve the detection limit of ctDNA, thereby achieving the purpose of improving the accuracy of MRD monitoring. Summary of the invention

[0013] In order to overcome the deficiencies of the above-mentioned prior art, the present invention provides a method, electronic device and storage medium for detecting tiny residual lesions based on blood ctDNA, which can monitor the MRD status of tumor patients at a relatively low cost, and the sensitivity of ctDNA detection reaches one in ten thousand.

[0014] In order to achieve the above purpose, this application adopts the following technical solutions:

[0015] In a first aspect, the present invention provides a method for detecting minimal residual lesions based on blood ctDNA, the detection method comprising the following steps:

[0016] S1, panel sequencing data: The cfDNA of the target region is amplified by primer binding, and the personalized panel sequencing data of the tumor patient is obtained by sequencing;

[0017] S2, automatically analyze the sequencing data. The specific analysis method is as follows:

[0018] S21, input the location information of the tumor patient's MRD monitoring site in the genome and the patient's personalized panel sequencing data;

[0019] S22, use fastp software to perform data preprocessing and analysis on the sequencing data, including removing sequencing adapters and low-quality sequences;

[0020] S23, using fgbio to read the UMI sequence of the pre-processed sequencing data, and using bwa software to compare and analyze the sequencing data after reading the UMI sequence with the hg19 genome;

[0021] S24, using fgbio to group the reads mapped to the genome by UMI, and divide the UMI into two groups according to different screening thresholds, one group is high-quality sequencing sequences with SSCS≥10, and the other group is standard quality sequencing sequences with SSCS≥3;

[0022] S25, high-quality sequencing sequences with SSCS ≥ 10 were subjected to mutation analysis, the detected mutation sites and the mutation sites monitored by MRD were taken as credible mutation sites, and standard-quality sequencing sequences with SSCS ≥ 3 were retrieved;

[0023] S26, removal of non-tumor-associated somatic mutations;

[0024] S3, MRD positive judgment:

[0025] When the average mutation rate of the MRD monitoring site is significantly higher than the background mutation rate, the MRD result is considered positive, otherwise it is negative;

[0026] S4, stores patient information and MRD results:

[0027] The MySQL database management system is used to store the patient's basic information, as well as the monitoring time and results of this MRD, to form a continuous monitoring record of the patient, to track the patient and make a prognostic judgment.

[0028] In the above technical solution, in S25, the criteria for recovering the standard quality sequencing sequences with SSCS ≥ 3 are as follows:

[0029] (1) If the sequence does not contain any mutated bases, retrieve the sequence;

[0030] (2) If there is a mutant base in the sequence that is consistent with the credible mutation site, the mutant base is retained;

[0031] (3) If there is a mutant base in the sequence that is inconsistent with the credible mutation site, the mutant base is set to N;

[0032] (4) If the number of N bases in the sequence exceeds 5, the sequence is discarded; otherwise, the sequence is retained.

[0033] In the above technical solution, the specific method for removing non-tumor related somatic mutations in S26 is:

[0034] K-means clustering analysis was used to perform unsupervised clustering of the mutation ratios and mutation quality values ​​of all mutation sites except the MRD monitoring sites. The clustering results were divided into two categories: one was background mutations with a low mutation ratio, and the other was somatic mutations with a high mutation ratio and high mutation quality. The second type of mutations was removed.

[0035] In the above technical solution, in S3, a t-test is used to perform a significance test on the mutation ratio of the MRD monitoring site and the mutation ratio of the background mutation. The specific steps are as follows:

[0036] S31, establish hypothesis and determine the test level α (0.05),

[0037] H0: The mutation ratio of the MRD monitoring site is equal to the mutation ratio of the background mutation.

[0038] H1: The mutation ratio of the MRD monitoring site is greater than the mutation ratio of the background mutation;

[0039] S32, calculate the test statistic,

[0040]

[0041] in,

[0042]

[0043]

[0044] X1 represents the average mutation value of the MRD monitoring site, n1 represents the number of mutations in the MRD monitoring site, and S1 represents the standard deviation of the mutation ratio in the MRD monitoring site.

[0045] X2 represents the average mutation rate of background mutations, n2 represents the number of mutations in the background mutations, and S2 represents the standard deviation of the proportion of mutations in the background mutations;

[0046] S33, calculate the P value and judge the significance,

[0047] If the P value is less than 0.05, the null hypothesis H0 is rejected and the alternative hypothesis H1 is accepted, that is, the mutation ratio of the MRD monitoring site is greater than the mutation ratio of the background mutation.

[0048] If the P value is greater than 0.05, the null hypothesis H0 is accepted, that is, the mutation ratio of the MRD detection site is no different from the mutation ratio of the background mutation.

[0049] In the above technical scheme, the test is carried out on an electronic device, which is equipped with an MRD monitoring data processing automation platform, and the installed software includes R, perl, python, Java, fastp, gatk, VarDict, Annovar, bedtools, samtools, picard, nextflow, bcftools, bwa and fgbio; the installed python packages include pandas and shutil; the downloaded databases include the human genome sequence hg19 version, the annotated gtf file of the human genome hg19 and the dbsnp database corresponding to hg19.

[0050] In a second aspect, the present invention provides an electronic device, including a processor and a memory, wherein the processor and the memory are connected, the memory is used to store a computer program, and the processor is used to call the computer program to execute the above-mentioned detection method.

[0051] In a third aspect, the present invention provides a computer storage medium, wherein the computer storage medium stores a computer program, wherein the computer program includes program instructions, and when the program instructions are executed by a processor, the above-mentioned detection method is executed.

[0052] The beneficial effects of the present invention are:

[0053] 1. Simple input, clear output, and wide application range. You only need to determine the mutation site to be monitored to achieve a fully automated analysis process from sequencing sequence to MRD diagnosis, which is applicable to all cfDNA MRD monitoring.

[0054] 2. There is no need to pair samples. Somatic mutations can be automatically identified through the characteristic variables of the data, and the mutation sites can be removed to avoid affecting the subsequent analysis of tumor mutation sites.

[0055] 3. MRD monitoring has low cost and high accuracy. Only 0.5G sequencing data is needed to complete MRD monitoring for 25 mutation sites.

[0056] 4. Establish continuous MRD monitoring and recording for individual patients, so as to facilitate the monitoring of the patient's disease progression at any time. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 is a schematic diagram of an electronic device of the present invention;

[0058] Figure 2 It is a flow chart of MRD analysis of the present invention. DETAILED DESCRIPTION

[0059] In order to better illustrate the purpose, technical solutions and advantages of the present invention, the present invention will be further described below in conjunction with specific embodiments. The present invention can be implemented in many different forms and should not be construed as being limited to the embodiments set forth herein. On the contrary, these embodiments are provided so that this disclosure will be thorough and complete, and the concept of the present invention will be fully conveyed to those skilled in the art, and the present invention will only be limited by the claims.

[0060] The present invention provides a method for detecting micro-residual lesions based on blood ctDNA, and the specific steps are as follows:

[0061] S1, personalized panel design and sequencing

[0062] Through the WES (whole exome sequencing) panel (probe combination or primer combination to capture the target region sequence) or a customized large panel for a certain type of tumor (the target region is greater than one million bases), the patient's genome is sequenced before tumor treatment. Through strict filtering and professional medical interpretation, cancer-related mutation sites are selected as MRD monitoring sites for personalized panel design, and a plasma cfDNA library containing a single-end UMI (unique molecular identifier) ​​adapter and primer sequence is constructed. The cfDNA in the target region is amplified by primer binding, and sequencing is performed on the machine to obtain the sequencing data of the personalized panel of tumor patients.

[0063] S2 is a fully automated device for processing sequencing data of personalized panels of tumor patients. The specific process is as follows:

[0064] S21, construct electronic equipment and corresponding analysis platform. Figure 1 As shown. The corresponding software needs to be installed on the electronic device to form a bioinformatics analysis platform. The main software that needs to be installed are: R, perl, python, Java, fastp, gatk, VarDict, Annovar, bedtools, samtools, picard, nextflow, bcftools, bwa and fgbio. The python packages that need to be installed are: pandas and shutil. The databases that need to be downloaded are: human genome sequence hg19 version, human genome hg19 annotation gtf (General Transfer Format) file and hg19 corresponding dbsnp database. The software and database can be downloaded and installed once, without repeated installation.

[0065] The MRD monitoring data processing automation platform of the present invention, which is constructed by building various databases on electronic equipment and cooperating with various software and hardware, is compatible with MRD monitoring data analysis of various types of cancer.

[0066] S22, automated analysis of sequencing data from personalized panels for cancer patients

[0067] S221, input the location information of the tumor patient's MRD monitoring site in the genome and the sequencing data of the patient's personalized panel.

[0068] S222, use fastp software to perform data preprocessing and analysis on the sequencing data, including removing sequencing adapters and low-quality sequences.

[0069] S223, use fgbio to read the UMI (unique molecular identifier) ​​sequence of the pre-processed sequencing data, and use bwa software to compare the sequencing data after reading the UMI sequence with the hg19 genome

[0070] S224, using fgbio to group the reads mapped to the genome by UMI, and further divide the UMI into two groups according to different screening thresholds, one group is high-quality sequencing sequences with SSCS (single-strand consensus sequence) ≥ 10, and the other group is standard quality sequencing sequences with SSCS ≥ 3.

[0071] S225, the key to improving the sensitivity of ctDNA detection lies in improving the accuracy of detection under ultra-low frequency mutations of single sites.

[0072] Since sequencing itself will produce base errors, and operations such as PCR amplification during the experiment will also produce technical base errors, this will inevitably lead to systematic base errors in the sequencing data itself. These base errors are not tumor-related base mutations. We collectively call them background noise mutations. These noise mutations will interfere with the detection of true tumor mutations during ultra-low frequency mutation detection.

[0073] For example, when we detect a base mutation at a certain monitoring site in the sequencing data of an MRD sample, due to the very low mutation frequency, such as a mutation ratio of 5 in 10,000, the number of reads supporting the mutation may be only 2. Then, we cannot accurately determine whether the mutation at this site is a true tumor-related base mutation or a background error caused by the sequencing process.

[0074] Therefore, the key to improving the detection accuracy of ultra-low-frequency mutations at single sites is to reduce the background error rate and ensure that the mutations detected in the sequencing data are tumor-related mutations. In other words, the average mutation frequency of the monitored tumor-related sites is significantly higher than the noise mutation frequency. The following is the specific implementation process and algorithm:

[0075] 1. Reduce background error rate: Perform mutation analysis on high-quality sequences with SSCS≥10, and use the detected mutation sites and the mutation sites monitored by MRD as credible mutation sites. In order to improve the effective utilization of the template chain and ensure the coverage depth of a single mutation site, retrieve the standard quality sequencing sequences with SSCS≥3. The sequencing sequence retrieval standards are as follows:

[0076] (1) If the sequence does not contain any mutated bases, retrieve the sequence;

[0077] (2) If there is a mutant base in the sequence that is consistent with the credible mutation site, the mutant base is retained;

[0078] (3) If there is a mutant base in the sequence that is inconsistent with the credible mutation site, the mutant base is set to N;

[0079] (4) If the number of N bases in the sequence exceeds 5, the sequence is discarded, otherwise the sequence is retained;

[0080] The method of reducing the background error rate is to set different thresholds through UMI tags to screen high-quality sequences and standard-quality sequences, use high-quality sequences to obtain high-confidence mutation sites, and retrieve sequences consistent with high-confidence mutation sites and sequences without mutations in standard-quality sequences. This method can maintain a high effective sequencing depth under the condition of low sequencing amount, saving sequencing costs, and significantly reducing the background error rate.

[0081] 2. Removal of non-tumor related somatic mutations including clonal hematopoietic mutations

[0082] Clonal hematopoietic mutations refer to somatic mutations that occur in human cells throughout their life, some of which can drive clonal expansion. These mutations are somatic mutations, not background mutations, and do not change with the progression of tumor development. Therefore, before making an MRD-positive judgment, these mutations need to be filtered out to remove their impact on background mutations. The specific removal principle is as follows:

[0083] After UMI correction, the proportion of background mutations is low, and the mutations generally conform to the normal distribution. Clonal hematopoietic mutations are true somatic mutations, with a higher mutation ratio than background mutations and do not change with the progression of the tumor disease. Therefore, based on this feature, we use k-means clustering analysis to perform unsupervised clustering of the mutation ratios and mutation quality values ​​of all mutation sites except the MRD monitoring sites. The clustering results are divided into two categories, one is background mutations with a low mutation ratio, and the other is somatic mutations with a high mutation ratio and high mutation quality (including clonal hematopoietic mutations). The second type of mutations are removed, thereby removing clonal hematopoietic mutations.

[0084] The method of removing non-tumor related somatic mutations including clonal hematopoietic mutations. The mutation ratio of clonal hematopoietic mutations and conventional human SNP sites will not change with the progression of the tumor, and is quite different from the mutation ratio of tumor monitoring sites. Therefore, the k-means dynamic programming clustering method is used to utilize pure mathematical features to automatically filter somatic mutations including clonal hematopoietic mutations, thereby reducing interference with tumor mutation sites.

[0085] S3, MRD positive judgment:

[0086] When the average mutation rate of the MRD monitoring site is significantly higher than the background error rate, it means that the mutation of the MRD monitoring site is not a random background mutation, but a real mutation caused by the tumor. At this time, we can consider the MRD result to be positive, otherwise it is negative.

[0087] Since background mutations can be considered as randomly occurring erroneous bases, the mutation ratio of random erroneous bases in background mutations can be considered to conform to the normal distribution. The MRD monitoring sites reflect the residual tumor load intensity, so the mutation ratio of each mutation site is theoretically consistent. Due to the randomness of sequencing, the mutation ratio will fluctuate within a small range, but the overall mutation ratio of the MRD monitoring sites also conforms to the normal distribution.

[0088] Therefore, we can use the t-test to perform a significance test on the mutation ratio of the MRD monitoring site and the background mutation ratio. The specific steps are as follows:

[0089] 1. Establish hypothesis and determine the test level α (0.05)

[0090] H0: The mutation ratio of MRD monitoring sites is equal to the mutation ratio of background mutations

[0091] H1: The mutation ratio of MRD monitoring sites is greater than the mutation ratio of background mutations

[0092] 2. Calculate the test statistic

[0093]

[0094] in

[0095]

[0096]

[0097] X1 represents the average mutation value of the MRD monitoring site, n1 represents the number of mutations in the MRD monitoring site, and S1 represents the standard deviation of the mutation ratio in the MRD monitoring site.

[0098] X2 represents the average mutation rate of background mutations, n2 represents the number of mutations in the background mutations, and S2 represents the standard deviation of the mutation rate of mutations in the background mutations.

[0099] 3. Calculate P value and judge significance

[0100] If the P value is less than 0.05, the null hypothesis H0 is rejected and the alternative hypothesis H1 is accepted, that is, the mutation ratio of the MRD monitoring site is greater than the mutation ratio of the background mutation. If the P value is greater than 0.05, the null hypothesis H0 is accepted, that is, the mutation ratio of the MRD detection site is no different from the mutation ratio of the background mutation.

[0101] S4, stores patient information and MRD results

[0102] The MySQL database management system is used to store the patient's basic information, including name, age, gender, ID number or unique patient ID, as well as the monitoring time and results of this MRD, to form a continuous monitoring record of the patient, to track the patient and make a prognosis judgment.

[0103] Example 1

[0104] MRD monitoring is performed on mixed cell line samples. The analysis process is as follows: Figure 2 The MRD analysis flow chart is shown below:

[0105] 1. Screen 27 homozygous germline loci in the cell line as cancer monitoring loci, use normal human gDNA to dilute the cell line samples, and dilute the mutation ratios of the homozygous germline loci of the cell line to 0.05%, 0.02%, and 0.01% respectively to simulate the ultra-low frequency tumor mutation state in cfDNA. Set up 2 groups of repeated samples for each mutation ratio. At the same time, use normal human gDNA as the control group, and also set up two groups of repeats. A total of 8 groups of data were used for MRD monitoring analysis.

[0106] 2. Simulate an identity information form (name, gender, age, ID number) and input it into the computing device.

[0107] 3. These samples were subjected to multiplex PCR library construction and sequencing.

[0108] 4. The electronic device reads the sequencing data from the sequencer, and pre-processes the sequencing data using the software fastp to remove redundant adapters and low-quality sequencing sequences. The sequencing sequences with a sequence length of less than 30 bp after removing the adapters are removed.

[0109] 5. The filtered high-quality sequencing sequences are used by the CPU resources in the computing device, and the UMI information in the sequencing sequences is extracted using the fastq2bam command in the fgbio software. The information is stored in the file result in the bam (Binary Alignment Map) format.

[0110] 6. Use picard software to convert the bam format sequencing data containing UMI information into fastq format, and use bwa software to compare the sequencing sequence to the human reference genome (hg19 version), generate bam format alignment information, and use samtools to create an index file for the bam file.

[0111] 7. Use the GroupReadsByUmi command of the fgbio software to group the sequences using UMI information and aligned position information.

[0112] 8. Use CallMolecularConsensusReads of fgbio software to filter the UMI grouped comparison file to obtain the consensus sequence file. The filtering condition is that the lowest consistent UMI sequence is one, and the bases with a base quality value of less than 30 at a single site are marked as N. Other filtering conditions remain as the software default conditions.

[0113] 9. Use FilterConsensusReads of fgbio software to filter the alignment file of the consensus sequence to obtain high-quality SSCS sequences. The filtering thresholds are: the number of reads with the same UMI and the same alignment position ≥ 10, the quality value of a single base > 30, the overall error rate of reads ≤ 5%, the error rate of sequencing at a single site ≤ 5%, and the proportion of bases marked as N on the entire read ≤ 5%.

[0114] 10. Use VarDict software to detect mutations in high-quality SSCS sequences. Remove indel (insertion or deletion mutation) mutations, retain only SNV (singlenucleotide variants) mutations, store all mutation site information, and use self-written scripts to filter out non-tumor-related somatic mutation sites with high mutation frequencies through the k-means dynamic clustering method.

[0115] 11. Use FilterConsensusReads of fgbio software to filter the alignment file of the consensus sequence to obtain the SSCS sequence of standard quality. The filtering threshold is: the number of reads with the same UMI and the same alignment position ≥ 3, the quality value of a single base > 30, the overall error rate of reads ≤ 5%, the error rate of sequencing at a single site ≤ 5%, and the proportion of bases marked as N in the entire read ≤ 5%.

[0116] 12. According to the 27 cell line germline mutation sites and the high-quality mutation site information obtained in step 9, the entry set A is formed. Using a self-written script, the sequences with the same mutations as those in set A and the sequences without any mutations are retrieved from the standard quality SSCS sequences obtained in step 10. If a mutation site is inconsistent with the site in set A, the base of the site is set to N, that is, the site is skipped in subsequent analysis.

[0117] 13. The self-written MRD detection script was used to calculate the P value of the mutation ratio of the 27 target mutation sites equal to the background mutation ratio through the t-test. If the P value is <0.05, it is considered that the mutation ratio of the 27 target mutation sites is significantly higher than the background mutation ratio, that is, the MRD monitoring result of the sample is positive; if the P value is ≥0.05, it is considered that the mutation ratio of the 27 target mutation sites is not significantly higher than the background mutation ratio, that is, the MRD monitoring result of the sample is negative.

[0118] 14. Use MySQL to record the monitoring results and monitoring date into the sample information to facilitate subsequent continuous monitoring.

[0119] 15. Case statistics results:

[0120] Table 1 Sample MRD monitoring results

[0121]

[0122] It can be seen that the ratio of monitored mutations to expected mutations in all samples is highly consistent, which shows the accuracy of mutation ratio detection, and the MRD results are completely consistent with expectations, with an accuracy rate of 100%, indicating that the mutation detection limit of this method is less than one in ten thousand.

[0123] Example 2

[0124] Samples from 4 B-cell lymphoma patients and cfDNA samples from 4 normal subjects were collected for MRD monitoring and analysis. The specific steps are as follows:

[0125] 1. Based on the initial diagnosis mutation sites of each B-cell lymphoma patient, tumor-related mutation sites were screened out through mutation quality screening and report interpretation experts. The following tumor monitoring mutation sites were obtained for the 4 tumor patients:

[0126] Table 2 Number of tumor monitoring mutation sites in 4 tumor patients

[0127] sample Tumor1 Tumor2 Tumor3 Tumor4 Monitoring mutation number 20 22 23 25

[0128] Four customized panels were designed for MRD monitoring. Subsequently, blood from four patients after treatment and four normal subjects was extracted for high-throughput sequencing to form four sets of data, each corresponding to a customized panel for a tumor patient.

[0129] 2. Enter the identity information form (name, gender, age, ID number) of all samples into the electronic device.

[0130] 3. These 8 samples were subjected to cfDNA extraction, multiplex PCR library construction and sequencing.

[0131] 4. The electronic device reads the sequencing data from the sequencer and performs data preprocessing on the sequencing data through the software fastp to remove redundant adapter sequences, low-quality sequencing sequences, and sequencing reads with a sequence length of less than 30bp after removing the adapters.

[0132] 5. The filtered high-quality sequencing sequences are used through the CPU resources in the computing device, and the fastq2bam command in the fgbio software is used to extract the UMI information in the sequencing sequence, and the information is stored in the bam format file result.

[0133] 6. Use picard software to convert the bam format sequencing data containing UMI information into fastq format, and use bwa software to compare the sequencing sequence to the human reference genome (hg19 version), generate bam format alignment information, and use samtools to create an index file for the bam file.

[0134] 7. Use the GroupReadsByUmi command of the fgbio software to group the sequences using UMI information and aligned position information.

[0135] 8. Use CallMolecularConsensusReads of fgbio software to filter the UMI grouped comparison file to obtain the consensus sequence file. The filtering condition is that the lowest consistent UMI sequence is one, and the bases with a base quality value of less than 30 at a single site are marked as N. Other filtering conditions remain as the software default conditions.

[0136] 9. Use FilterConsensusReads of fgbio software to filter the alignment file of the consensus sequence to obtain high-quality SSCS sequences. The filtering thresholds are: the number of reads with the same UMI and the same alignment position ≥ 10, the quality value of a single base > 30, the overall error rate of reads ≤ 5%, the error rate of sequencing at a single site ≤ 5%, and the proportion of bases marked as N on the entire read ≤ 5%.

[0137] 10. Use VarDict software to detect mutations in high-quality SSCS sequences. Remove indel mutations and keep only SNV mutations. Store all mutation site information and use a self-written script to filter out somatic mutation sites with high mutation frequency using the k-means dynamic clustering method.

[0138] 11. Use FilterConsensusReads of fgbio software to filter the alignment file of the consensus sequence to obtain the SSCS sequence of standard quality. The filtering threshold is: the number of reads with the same UMI and the same alignment position ≥ 3, the quality value of a single base > 30, the overall error rate of reads ≤ 5%, the error rate of sequencing at a single site ≤ 5%, and the proportion of bases marked as N on the entire read ≤ 5%.

[0139] 12. According to the 27 cell line germline mutation sites and the high-quality mutation site information obtained in step 9, the compilation set A is formed. Using a self-written script, the sequences with the same mutations as those in set A and the sequences without any mutations are retrieved from the standard quality SSCS sequences obtained in step 11. If a mutation site is inconsistent with the site in set A, the base of the site is set to N, that is, the site is skipped in subsequent analysis.

[0140] 13. The self-written MRD detection script was used to calculate the P value of the mutation ratio of the 27 target mutation sites equal to the background mutation ratio through the t-test. If the P value is <0.05, it is considered that the mutation ratio of the 27 target mutation sites is significantly higher than the background mutation ratio, that is, the MRD monitoring result of the sample is positive; if the P value is ≥0.05, it is considered that the mutation ratio of the 27 target mutation sites is not significantly higher than the background mutation ratio, that is, the MRD monitoring result of the sample is negative.

[0141] 14. Use MySQL to record the monitoring results and monitoring date into the sample information to facilitate subsequent continuous monitoring.

[0142] 15. Case statistics results:

[0143] Table 3 Sample MRD monitoring results

[0144]

[0145]

[0146] From this we can see that: among the 4 tumor samples after treatment, 2 still showed MRD positivity, which means that the tumor has not been completely relieved and requires continued treatment; 2 showed MRD negativity, which means that the prognosis of tumor treatment is good; the other 4 normal control samples all showed MRD negativity, reflecting the accuracy of the algorithm.

[0147] Obviously, the above embodiments are merely examples for the purpose of clear explanation, and are not intended to limit the implementation methods. For those skilled in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to list all the implementation methods here. The obvious changes or modifications derived therefrom are still within the scope of protection of the invention.

Claims

1. An electronic device, characterized in that: The device comprises a processor and a memory, wherein the processor and the memory are connected, the memory is used to store a computer program, and the processor is used to call the computer program to execute a method for detecting micro-residual lesions based on blood ctDNA; The detection method comprises the following steps: S1, panel sequencing data: The cfDNA of the target region is amplified by primer binding, and the personalized panel sequencing data of the tumor patient is obtained by sequencing; S2, automatically analyze the sequencing data. The specific analysis method is as follows: S21, input the location information of the tumor patient's MRD monitoring site in the genome and the patient's personalized panel sequencing data; S22, use fastp software to perform data preprocessing and analysis on the sequencing data, including removing sequencing adapters and low-quality sequences; S23, using fgbio to read the UMI sequence of the pre-processed sequencing data, and using bwa software to compare and analyze the sequencing data after reading the UMI sequence with the hg19 genome; S24, using fgbio to group the reads mapped to the genome by UMI, and divide the UMI into two groups according to different screening thresholds, one group is high-quality sequencing sequences with SSCS≥10, and the other group is standard quality sequencing sequences with SSCS≥3; S25, high-quality sequencing sequences with SSCS ≥ 10 were subjected to mutation analysis, the detected mutation sites and the mutation sites monitored by MRD were taken as credible mutation sites, and standard-quality sequencing sequences with SSCS ≥ 3 were retrieved; S26, removal of non-tumor-associated somatic mutations; S3, MRD positive judgment: When the average mutation rate of the MRD monitoring site is significantly higher than the background mutation rate, the MRD result is considered positive, otherwise it is negative; S4, stores patient information and MRD results: The MySQL database management system is used to store the patient's basic information, as well as the monitoring time and results of this MRD, to form a continuous monitoring record of the patient, to track the patient and make a prognostic judgment; In S3, a t-test was used to perform a significance test on the mutation ratio of the MRD monitoring site and the mutation ratio of the background mutation. The specific steps are as follows: S31, establish hypothesis and determine test level α: 0.05, H0: The mutation ratio of the MRD monitoring site is equal to the mutation ratio of the background mutation. H1: The mutation ratio of MRD monitoring sites is greater than that of background mutations; S32, calculate the test statistic, in, X1 represents the average mutation value of the MRD monitoring site, n1 represents the number of mutations in the MRD monitoring site, and S1 represents the standard deviation of the mutation ratio in the MRD monitoring site. X2 represents the average mutation rate of background mutations, n2 represents the number of mutations in the background mutations, and S2 represents the standard deviation of the proportion of mutations in the background mutations; S33, calculate the P value and judge the significance, If the P value is less than 0.05, the null hypothesis H0 is rejected and the alternative hypothesis H1 is accepted, that is, the mutation ratio of the MRD monitoring site is greater than the mutation ratio of the background mutation. If the P value is greater than 0.05, the null hypothesis H0 is accepted, that is, the mutation ratio of the MRD detection site is no different from the mutation ratio of the background mutation.

2. The electronic device according to claim 1, characterized in that: In S25, the criteria for recovering standard quality sequencing sequences with SSCS ≥ 3 are as follows: (1) If the sequence does not have any mutated bases, retrieve the sequence; (2) If there is a mutant base in the sequence that is consistent with the credible mutation site, the mutant base is retained; (3) If there is a mutant base in the sequence that is inconsistent with the credible mutation site, the mutant base is set to N; (4) If the number of N bases in the sequence exceeds 5, the sequence is discarded; otherwise, the sequence is retained.

3. The electronic device according to claim 1, characterized in that: The specific method for removing non-tumor related somatic mutations in S26 is: K-means clustering analysis was used to perform unsupervised clustering of the mutation ratios and mutation quality values ​​of all mutation sites except the MRD monitoring sites. The clustering results were divided into two categories: one was background mutations with a low mutation ratio, and the other was somatic mutations with a high mutation ratio and high mutation quality. The second type of mutations was removed.

4. The electronic device according to claim 1, characterized in that: The detection method is carried out on an electronic device, which is equipped with an MRD monitoring data processing automation platform. The installed software includes R, perl, python, Java, fastp, gatk, VarDict, Annovar, bedtools, samtools, picard, nextflow, bcftools, bwa and fgbio; the installed python packages include pandas and shutil; the downloaded databases include the hg19 version of the human genome sequence, the annotated gtf file of the human genome hg19 and the dbsnp database corresponding to hg19.

5. A computer storage medium, characterized in that: The computer storage medium stores a computer program, wherein the computer program includes program instructions, and when the program instructions are executed by the processor, a method for detecting micro-residual lesions based on blood ctDNA is executed; The detection method comprises the following steps: S1, panel sequencing data: The cfDNA of the target region is amplified by primer binding, and the personalized panel sequencing data of the tumor patient is obtained by sequencing; S2, automatically analyze the sequencing data. The specific analysis method is as follows: S21, input the location information of the tumor patient's MRD monitoring site in the genome and the patient's personalized panel sequencing data; S22, use fastp software to perform data preprocessing and analysis on the sequencing data, including removing sequencing adapters and low-quality sequences; S23, using fgbio to read the UMI sequence of the pre-processed sequencing data, and using bwa software to compare and analyze the sequencing data after reading the UMI sequence with the hg19 genome; S24, using fgbio to group the reads mapped to the genome by UMI, and divide the UMI into two groups according to different screening thresholds, one group is high-quality sequencing sequences with SSCS≥10, and the other group is standard quality sequencing sequences with SSCS≥3; S25, high-quality sequencing sequences with SSCS ≥ 10 were subjected to mutation analysis, the detected mutation sites and the mutation sites monitored by MRD were used as credible mutation sites, and standard-quality sequencing sequences with SSCS ≥ 3 were retrieved; S26, removal of non-tumor-associated somatic mutations; S3, MRD positive judgment: When the average mutation rate of the MRD monitoring site is significantly higher than the background mutation rate, the MRD result is considered positive, otherwise it is negative; S4, stores patient information and MRD results: The MySQL database management system is used to store the patient's basic information, as well as the monitoring time and results of this MRD, to form a continuous monitoring record of the patient, to track the patient and make a prognostic judgment; In S3, a t-test was used to perform a significance test on the mutation ratio of the MRD monitoring site and the mutation ratio of the background mutation. The specific steps are as follows: S31, establish hypothesis and determine test level α: 0.05, H0: The mutation ratio of the MRD monitoring site is equal to the mutation ratio of the background mutation. H1: The mutation ratio of MRD monitoring sites is greater than that of background mutations; S32, calculate the test statistic, in, X1 represents the average mutation value of the MRD monitoring site, n1 represents the number of mutations in the MRD monitoring site, and S1 represents the standard deviation of the mutation ratio in the MRD monitoring site. X2 represents the average mutation rate of background mutations, n2 represents the number of mutations in the background mutations, and S2 represents the standard deviation of the proportion of mutations in the background mutations; S33, calculate the P value and judge the significance, If the P value is less than 0.05, the null hypothesis H0 is rejected and the alternative hypothesis H1 is accepted, that is, the mutation ratio of the MRD monitoring site is greater than the mutation ratio of the background mutation. If the P value is greater than 0.05, the null hypothesis H0 is accepted, that is, the mutation ratio of the MRD detection site is no different from the mutation ratio of the background mutation.

6. The computer storage medium according to claim 5, characterized in that: In S25, the criteria for recovering standard quality sequencing sequences with SSCS ≥ 3 are as follows: (1) If the sequence does not have any mutated bases, retrieve the sequence; (2) If there is a mutant base in the sequence that is consistent with the credible mutation site, the mutant base is retained; (3) If there is a mutant base in the sequence that is inconsistent with the credible mutation site, the mutant base is set to N; (4) If the number of N bases in the sequence exceeds 5, the sequence is discarded; otherwise, the sequence is retained.

7. The computer storage medium according to claim 5, characterized in that: The specific method for removing non-tumor related somatic mutations in S26 is: K-means clustering analysis was used to perform unsupervised clustering of the mutation ratios and mutation quality values ​​of all mutation sites except the MRD monitoring sites. The clustering results were divided into two categories: one was background mutations with a low mutation ratio, and the other was somatic mutations with a high mutation ratio and high mutation quality. The second type of mutations was removed.

8. The computer storage medium according to claim 5, characterized in that: The detection method is carried out on an electronic device, which is equipped with an MRD monitoring data processing automation platform. The installed software includes R, perl, python, Java, fastp, gatk, VarDict, Annovar, bedtools, samtools, picard, nextflow, bcftools, bwa and fgbio; the installed python packages include pandas and shutil; the downloaded databases include the hg19 version of the human genome sequence, the annotated gtf file of the human genome hg19 and the dbsnp database corresponding to hg19.

Images