Ultra-high depth sequencing-based tiny residual focus detection method and system

Abstract

The invention discloses a tiny residual focus detection method and system based on ultra-high depth sequencing, and relates to the technical field of tiny residual focus intelligent detection.The tiny residual focus detection method comprises the following steps that on the basis of a sequencing library, splitting is conducted according to a sample index to obtain a to-be-detected sample, and a consensus sequence is obtained according to a molecular identifier of the to-be-detected sample; based on a consensus sequence, filtering out the consensus sequence of which the mass value is less than 25 or the failysize is less than 3, and combining a subtype type and a distance from a fragment edge as noise introduced into an original nucleic acid molecular chain; a context sequence (context) and a chain direction are used as noise for introducing the capture level of PCR amplification; based on the noise level, in combination with tumor priori knowledge, the ctDNA level is estimated, and the MRD state is determined by detecting the significance of a molecular signal source. According to the invention, the sensitivity and specificity of MRD detection are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent detection of minimal residual lesions, and in particular to a method and system for detecting minimal residual lesions based on ultra-high depth sequencing. Background Art

[0002] The important role of circulating tumor DNA (ctDNA) as a biomarker in the detection of minimal residual disease (MRD) in solid tumors has been demonstrated by more and more clinical cohorts. Through liquid biopsy technology and bioinformatics methods, the detection and identification of ctDNA signals can be used to predict the risk of recurrence after radical treatment of various solid tumors. However, there are still a series of technical and biological challenges in the clinical application of ctDNA-based MRD detection.

[0003] In MRD detection, since there are abundant sources of free DNA in the blood, often only trace amounts of DNA are released from tumors. At the same time, mutations from clonal hematopoietic sources, subclonal mutations caused by tumor heterogeneity, and mutations from other unknown sources will lead to low levels of true ctDNA signals and difficulty in identification. Low levels of ctDNA are an important factor causing false negatives in MRD detection. On the other hand, technical noise generated by sample processing and library sequencing, such as DNA chain damage during library preparation, damage during hybridization capture, amplification errors during PCR amplification, and sequencing errors during sequencing, are the main factors causing false positives in MRD detection.

[0004] There are two types of current MRD detection strategies. One is the tumor-agonostic strategy, which uses a whole-genome approach, does not rely on prior variation map information, and enriches signals through methylation or fragment omics. The other is the tumor-informed strategy, which simultaneously tracks genomic variations from multiple known tumor sources, identifies the corresponding variant genotypes, and then determines the ctDNA level. Although the tumor-agonostic strategy has fewer operational links for a single test, it has a wide range and low-depth technical route, requires complex bioinformatics algorithms and matching baseline models, and is difficult to generalize tumor-specific signals, and has insufficient specificity. The tumor-informed strategy, although it increases the reliance on tumor priors, has very high sensitivity and specificity by jointly tracking personalized variations after building a variation map.

[0005] At present, the MRD detection based on tumor priors has different focuses on the way to suppress background noise according to different technical routes. In the group customized panel, the background baseline model constructed by the healthy group is mainly used to distinguish the background noise through statistical tests; in the individual customized panel, the consensus base is mainly obtained by molecular tagging technology, and the upstream and downstream context trimers of the genome from its own source are used as the background, such as C[A>T]G, C[G>A]G, etc. for classification, and the background error rate in the sample is evaluated.

[0006] A background noise database is established using data from a healthy cohort. For each detected variant, the significance between the frequency in the plasma sample and the frequency in the normal population background database is tested to determine whether the variant may be tumor-derived, thereby eliminating false positive results. This approach ignores the differences between data samples, lacks a stratified analysis of technical noise sources, and the deviation noise caused by batch effects cannot be completely removed. Upgrading the library construction and sequencing process also requires a large cost to re-establish the background noise database as a baseline, which is time-consuming and labor-intensive.

[0007] Using the trimer of the genomic background context, the sample-specific background noise error rate is calculated, and the depth of the detection site and the support number division threshold are combined as the basis for site determination. This method is too general for background assessment, does not make full use of the data, lacks decomposition of the background noise source level, and is difficult to balance specificity and sensitivity. In addition, the background noise assessment is relatively separated from the model for enriching the real signal, and there is a lack of effective models for the determination of ctDNA in plasma, and there is also a lack of effective and explainable models for the quantification of ctDNA. Summary of the invention

[0008] In order to solve the above problems, the purpose of the present invention is to systematically identify and evaluate the noise introduced in each link of NGS through supersaturated sequencing under molecular tagging technology, and perform feature extraction and statistical modeling on background errors at different levels, while establishing a ctDNA evaluation model that suppresses noise and enriches signals, thereby improving the specificity and sensitivity of MRD detection.

[0009] In order to achieve the above technical objectives, the present application provides a method for detecting micro-residual lesions based on ultra-high depth sequencing, comprising the following steps: Based on the sequencing library, the samples to be tested are split according to the sample index, and the consensus sequence is obtained according to the molecular identifier of the sample to be tested; Based on the consensus sequence, by filtering out consensus sequences with a quality value less than 25 or a family size less than 3, the subtype type and the distance from the fragment edge are combined as the noise level introduced into the original nucleic acid molecule chain; the PCR level noise is evaluated by combining the context and chain specificity.

[0010] Based on the noise level and combined with prior knowledge of the tumor, the ctDNA level is estimated, and the MRD status is determined by testing the significance of the molecular signal source.

[0011] Preferably, in the process of obtaining the sequencing library, a sequencing adapter is added to the nucleic acid molecule by connecting the adapter, and the adapter-nucleic acid complex is amplified by polymerase chain reaction to establish a sequencing library, wherein the adapter is a molecular identifier.

[0012] Preferably, in the process of obtaining the molecular identifier of the sample to be tested, the adapter and low-quality sequencing reads of the sample to be tested are removed, and the molecular identifier is obtained after sequence alignment.

[0013] Preferably, in the process of obtaining the consensus sequence, the sequencing reads are grouped according to the molecular identifier and the position aligned to the reference genome, and the base sequences on the repeated sequencing reads are voted to form a consensus sequence.

[0014] Preferably, in the process of evaluating the noise level, the sequencing read pairs containing both the positive strand template and the negative strand template in the group are taken as the first consensus sequence duplex; the consensus sequence formed by the sequencing read pairs of only one side template strand, that is, only the sequencing reads containing the positive strand or the negative strand, is taken as the second consensus sequence simplex; The noise level was evaluated based on the first consensus sequence duplex and the second consensus sequence simple.

[0015] Preferably, in the process of evaluating the noise level, after excluding high-frequency (VAF>=1%) unknown mutations, tumor somatic mutation sites of known origin, clonal hematopoietic mutations, and germline mutations, the noise level introduced into the original nucleic acid molecule chain is evaluated based on the first consensus sequence duplex combined with the subtype type and the distance from the edge of the fragment. The second consensus sequence simplex is combined with the context and chain specificity to evaluate the noise at the PCR level.

[0016] Preferably, in the process of combining tumor prior knowledge, variants are identified in samples with a high tumor ratio (>=20%), and the somatic variant information of the tumor origin is determined by deducting germline variants and clonal hematopoietic variants from paired leukocyte control sample data; combined with the information of the prior samples, the somatic variants are classified, clustered and quantitatively sorted, the classification categories are major clones and subclones, and the quantitative indicators are copy number and variant abundance. The variants are filtered and sorted according to the abundance, and a traceable variant set is obtained to construct a variant map for tumor priors.

[0017] Preferably, in the process of obtaining a traceable variation set, single-base point mutations, short-fragment insertion or deletion mutations, and long-fragment structural mutations are used as traceable variation sets, wherein the noise of point mutations and short insertion or deletion mutations is modeled and evaluated through the characteristics of genomic background context, edge distance, and chain variation direction; the noise of structural variation fragments is modeled and evaluated through the characteristics of variation direction and split length.

[0018] Preferably, in the process of determining the MRD status, based on the depth information of the sample to be tested and according to the evaluation model for evaluating the first noise and evaluating the second noise, the ctDNA level is estimated by calculating the maximum likelihood probability of the number of supporting molecules under multiple variations, and the expected number of molecules for evaluating different circulating tumor proportions is obtained and the probability is calculated. The significance is determined by the likelihood ratio test, thereby determining the MRD status.

[0019] The present invention discloses a system for detecting micro residual lesions based on ultra-high depth sequencing, which is used to implement the above-mentioned method for detecting micro residual lesions based on ultra-high depth sequencing. The system comprises: A data processing unit is used to split the sequencing library according to the sample index to obtain the sample to be tested, and obtain the consensus sequence according to the molecular identifier of the sample to be tested; A noise level evaluation unit is used to filter out consensus sequences with a quality value less than 25 or a family size less than 3 based on the consensus sequence, and combine the subtype type and the distance from the edge of the fragment as the noise level introduced into the original nucleic acid molecule chain; The detection unit is used to estimate the ctDNA level based on the noise level and combined with the prior knowledge of the tumor, and determine the MRD status by testing the significance of the source of the molecular signal.

[0020] The present invention discloses the following technical effects: The present invention can distinguish the sources of noise in the NGS detection process and evaluate the characteristics and levels of noise introduced in different links. It can also be used to compare the noise levels introduced in different experimental links. For example, duplex level noise can reflect the degree of jagged ends and damage of cfDNA samples, and simplex level noise can reflect the noise level brought about by sample PCR amplification and capture processes. It can also be used to react to the performance of reagents in the library construction, amplification and capture process.

[0021] The information on copy number and clone status in addition to the allele frequency of tumor tissue can accurately reflect the tumor proportion of tumor tissue, and combined with the hierarchical noise model, it is used to analyze the significance of the signal in plasma samples originating from tumor tissue. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0023] Figure 1 Describes the main steps of NGS sequencing and the noise signals introduced by each step Figure 2 It is a schematic diagram of the method flow of the present invention. DETAILED DESCRIPTION

[0024] To make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. The components of the embodiments of the present application generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the application claimed for protection, but merely represents the selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without making creative work belong to the scope of protection of the present application.

[0025] like Figure 1-2 As shown, the present invention provides a method for detecting micro-residual lesions based on ultra-high depth sequencing, which uses molecular tags under ultra-high depth sequencing to distinguish the background noise introduced by different NGS links, characterizes the background error rate of each link source according to the source extraction characteristics, estimates the noise level of the sample itself, and develops an algorithm process to eliminate false positive signals without baseline dependence, maximize the accuracy of determining the ctDNA status, and then determine the ctDNA level. The main process includes the following steps: 1. Sequencing library preparation and sequencing: Prepare sequencing libraries. During library preparation, sequencing adapters are added to nucleic acid molecules (e.g., DNA molecules) by connecting adapters, which include unique molecular identifiers (UMIs). After the adapters are connected, the adapter-nucleic acid complex is amplified by polymerase chain reaction (PCR). During PCR amplification, UMIs are replicated along with the attached DNA fragments, so that sequencing reads from the same original fragment can be identified, and can also be used to distinguish noise levels from different sources of NGS. After the library is completed, enrichment, hybridization, and amplification are performed, and finally provided to the sequencer for sequencing.

[0026] 2. Off-line data preprocessing and comparison: The sequencing data were split according to the sample index using the bcl2fastq software to obtain the FASTQ file of the sample to be tested, and then the fastp software was used to remove the adapters and low-quality reads to obtain the filtered FASTQ file; the genome alignment tool BWA mem -M was used to align the sequences to obtain the bam file.

[0027] 3. Obtain consensus sequence through molecular tags Use the pysam module to read the bam file, group the reads according to the UMI tag and the position aligned to the reference genome, and the number of repeated sequencing reads contained in the group is called the family size of the sequence. The base sequences on the repeated reads form a consensus sequence by voting. These reads are replicated during the experimental and sequencing stages, so the consensus sequence also represents the original nucleic acid fragments from which these reads originate. If the group contains reads pairs derived from both the positive and negative strand templates, such a consensus sequence is called a duplex. If the consensus sequence is formed by reads pairs on only one side of the template chain, that is, it only contains reads on the positive or negative strand, then such a consensus sequence is called a simplex.

[0028] 4. Estimation of noise levels at different NGS stages through hierarchical features: First, in the process of evaluating noise, we exclude high-frequency (VAF>=1%) unknown variants, tumor somatic variant sites of known origin, clonal hematopoietic variants, and germline variants. The sequencing base quality value reflects the error rate of the sequencing process. A base quality value of about 30 represents an error rate of 0.1%. In order to accurately evaluate the background noise below 0.1% and avoid the influence of sequencing batch deviation, we filter out consensus sequences with a quality value less than 25 or a family size less than 3. The consensus sequence may be consistent with the reference genome sequence or inconsistent. Inconsistent consensus sequences will be used as the basis for subsequent background evaluation.

[0029] Since UMI undergoes library amplification and capture enrichment after being connected to nucleic acid molecules, the base changes at the duplex level are derived from before PCR amplification, which may come from real mutations, or from adapter connection and DNA damage in the early stage of library construction. The base changes at the simplex level may include real mutations and errors in PCR amplification, enrichment and capture. The noise levels from different sources are different, and their noise levels and characteristics are also different, so it is necessary to construct a noise model in a hierarchical manner.

[0030] Studies have shown that most plasma DNA molecules need to be filled in during library construction because of their jagged ends. During this process, the synthesis bias of DNA polymerase and sample damage will cause base changes. The base changes produced in this process are on the original nucleic acid fragment molecules before PCR, and duplex can reflect their characteristics and levels. We combined subtype type and distance from the edge of the fragment as features to estimate the noise level introduced on the original nucleic acid molecule chain during sample extraction and library construction.

[0031] In the process of PCR amplification, i.e. cluster generation, the signal of the original template is "amplified", and errors are also introduced. Studies have shown that the types of errors introduced by PCR mainly include mismatches, insertions and deletions (indels), etc., and the probability of errors is different for different background bases upstream and downstream of the sequence. The upstream and downstream trimers of the bases, the type of mutation (subtype), and the chain direction (forward or reverse) all affect the probability of errors, so these factors are included as features to measure the noise level generated during PCR amplification and cluster generation.

[0032] 5. Estimating ctDNA levels by combining tumor priors and noise model tests Variants are identified in samples (usually surgical tissues) with a high tumor proportion (>=20%), and germline variants and clonal hematopoietic variants are deducted from paired leukocyte control sample data to determine the somatic variant information of the tumor origin. Then, combined with the information of the prior samples, the somatic variants are classified, clustered, and quantitatively sorted. The classification categories are primary clones and subclones, and the quantitative indicators are copy number and variant abundance. Differences in the proportion of tumor cells, tumor heterogeneity, and genomic copy number lead to differences in the abundance of detected variants. These variants are filtered and sorted according to their abundance, thereby selecting the top number of variants as a set of traceable variants to construct a priori variant map of the tumor.

[0033] Each variant information of the tumor prior map is standardized as a relative coefficient , and its calculation formula is as follows: in, It refers to the tumor cell proportion, that is, the proportion of tumor cells in the a priori sample; It refers to the circulating tumor ratio, that is, the proportion of tumor molecules in cell-free DNA (cfDNA).

[0034] Trackable mutations include single-base point mutations (SNVs), short insertion or deletion mutations (INDELs), and long structural mutations (SVs). The technical noise of point mutations and short insertion or deletion mutations mainly comes from different links of NGS. The changes in the base level on duplex and simplex molecules are modeled through features such as context, edge distance, and chain variation direction; the noise of structural variation fragments mainly comes from repeated sequences or random sequences, which causes the sequencing reads to be incorrectly mapped to different gene locations, and is modeled through variation direction and split length features.

[0035] Combining the deep information of sequencing data and the hierarchical background noise model, we evaluate different The expected number of molecules is as follows: ; in, is the expected number of molecules for the ith mutation based on the j-layer background noise model; is the sequencing depth of the j-layer background noise model corresponding to the ith variant; is the background error rate of the j-layer background noise model corresponding to the ith mutation.

[0036] Calculate the maximum likelihood probability of the number of supporting molecules under multiple mutations and obtain the estimated : .

[0037] 6. Check the significance of the molecular signal source: Through multiple tracking variations combined with the hierarchical background noise model, based on the H0 hypothesis that the signal comes from the background noise, we estimate The number of molecules in the case and the corresponding probability are calculated, based on the H1 hypothesis that the signal comes from the tumor, estimating The MRD status was determined by calculating the expected number of molecules in the case and calculating the probability, followed by a likelihood ratio test to determine the significance.

[0038] The formula for calculating the probability of H0 hypothesis is as follows: ; The formula for calculating the probability of H1 hypothesis is as follows: ; In the formula, y represents the molecular signal actually observed at the site; Y represents the molecular signal at the site The expected molecular signal under the conditions; j represents the three-level signal, namely the simplex level positive chain, the simplex level negative chain and the duplex level; X represents the level of each site at each level The expected molecular signal under the conditions of ; n represents the tracked site; P represents the probability based on the H1 hypothesis that the signal comes from the tumor.

[0039] The formula for the sample level significance test is as follows: .

[0040] The input files required by the present invention include: (1) Sequencing data of samples with a high tumor proportion (>=20%); (2) Sequencing data of control samples; (3) Sequencing data of plasma samples after surgery or treatment.

[0041] The output results of the present invention include: (1) Consensus sequence based on the same alignment position between the UMI tag and the reference genome; (2) Inconsistent sequences based on the same alignment position between the UMI tag and the reference genome; (3) Noise files (duplex level) based on position effects and different subtype features; (4) Noise files based on chain direction, context bases, and variation direction (simplex level); (5) Variant profile information files, including information on these variants in leukocytes, tissues, and plasma samples; (6) ctDNA levels estimated by combining a priori mutation profiles and hierarchical noise models; (7) The significance level of the test combined with the prior variation map and the hierarchical noise model.

[0042] The present invention proposes a stratified analysis based on the noise introduced at different stages of NGS data, and a comparative analysis of the noise sources and their characteristic levels, which is used for noise reduction and MRD detection in subsequent tracking sites. This is more detailed and accurate than the noise assessment methods of previous studies and can improve the sensitivity and specificity of MRD detection.

[0043] In addition, the present invention makes full use of information such as clonal evolution during tumor development. In addition to the allele frequency of tumor tissue, it also incorporates tumor mutation allele frequency, copy number and clonality information, which can more accurately estimate the proportion of ctDNA in plasma and improve the performance of MRD detection.

[0044] The present invention first uses plasma leukocytes to deduct false positive sites introduced by clonal hematopoiesis to improve the specificity of MRD detection. Then, a stratified analysis was performed according to the noise from different sources in the NGS sequencing process, which can more accurately reflect the level and characteristics of the noise source, and can eliminate the false positive rate caused by the introduction of noise in each link as much as possible without excessive loss of true positive signals. On this basis, combined with tumor tissue information, including mutant allele frequency, tumor proportion, copy number, and cloning status, it is assumed that the relevant mutation is a single copy in the tumor tissue and the tissue tumor accounts for 100%. The method can more accurately reflect the plasma tumor proportion. Combining multiple tracking sites with stratified background noise to test the significance level of the signal in the plasma originating from the tumor tissue has a very good effect on distinguishing the source of the signal in the plasma.

[0045] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0046] In the description of the present invention, it should be understood that the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0047] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A method for detecting minimal residual lesions based on ultra-high depth sequencing, characterized in that: The following steps are involved: Based on the sequencing library, the samples to be tested are obtained by splitting according to the sample index, and a consensus sequence is obtained according to the molecular identifier of the sample to be tested; Based on the consensus sequence, by filtering out consensus sequences with a quality value less than 25 or a family size less than 3, the subtype type and the distance from the edge of the fragment are combined as the noise level introduced into the original nucleic acid molecule chain; Based on the noise level and combined with prior knowledge of the tumor, the ctDNA level is estimated, and the MRD status is determined by examining the significance of the source of the molecular signal.

2. A method for detecting minimal residual lesions based on ultra-high depth sequencing according to claim 1, characterized in that: In the process of obtaining the sequencing library, a sequencing adapter is added to the nucleic acid molecule by connecting the adapter, and the adapter-nucleic acid complex is amplified by polymerase chain reaction to establish the sequencing library, wherein the adapter is a molecular identifier.

3. A method for detecting minimal residual lesions based on ultra-high depth sequencing according to claim 2, characterized in that: In the process of obtaining the molecular identifier of the sample to be tested, the connector and low-quality sequencing reads of the sample to be tested are removed, and the molecular identifier is obtained after sequence alignment.

4. A method for detecting minimal residual lesions based on ultra-high depth sequencing according to claim 3, characterized in that: In the process of obtaining the consensus sequence, the sequencing reads are grouped according to the molecular identifier and the position aligned to the reference genome, and the base sequences on the repeated sequencing reads are voted to form the consensus sequence.

5. A method for detecting minimal residual lesions based on ultra-high depth sequencing according to claim 4, characterized in that: In the process of evaluating the noise level, the sequencing read pairs containing both positive and negative strands in the group are taken as the first consensus sequence duplex; the consensus sequence is formed by sequencing read pairs of only one side template strand, that is, only containing sequencing reads of the positive strand or the negative strand, as the second consensus sequence simplex; The noise level is evaluated respectively according to the first consensus sequence duplex and the second consensus sequence simple.

6. A method for detecting minimal residual lesions based on ultra-high depth sequencing according to claim 5, characterized in that: In the process of evaluating the noise level, after excluding high-frequency (VAF>=1%) unknown variants, tumor somatic variant sites of known origin, clonal hematopoietic variants and germline variants, the noise level introduced into the original nucleic acid molecule chain was evaluated based on the first consensus sequence duplex combined with the subtype type and the distance from the fragment edge; the second consensus sequence simplex, combined with context and chain specificity, was used to evaluate the noise at the PCR level.

7. A method for detecting minimal residual lesions based on ultra-high depth sequencing according to claim 6, characterized in that: In the process of combining tumor prior knowledge, variants are identified in samples with a high tumor ratio (>=20%), and germline variants and clonal hematopoietic variants are deducted from the paired leukocyte control sample data to determine the somatic variant information of the tumor origin; combined with the information of the prior samples, the somatic variants are classified, clustered and quantitatively sorted. The classification categories are main clones and subclones, and the quantitative indicators are copy number and variant abundance. The variants are filtered and sorted according to the abundance to obtain a traceable variant set and construct a variant map for tumor priors.

8. A method for detecting minimal residual lesions based on ultra-high depth sequencing according to claim 7, characterized in that: In the process of obtaining the traceable variation set, single-base point mutations, short-fragment insertion or deletion mutations, and long-fragment structural mutations are taken as the traceable variation set, among which the noise on the original chain of point mutations and short insertion or deletion is estimated by subtype and edge distance level; the noise introduced by PCR amplification capture is modeled and evaluated based on the characteristics of context and chain variation direction.

9. A method for detecting minimal residual lesions based on ultra-high depth sequencing according to claim 8, characterized in that: In the process of determining the MRD status, based on the depth information of the sample to be tested and according to the evaluation model for evaluating the first noise and evaluating the second noise, the maximum likelihood probability of the number of supporting molecules under multiple variations is calculated, the ctDNA level is estimated, the expected number of molecules for evaluating different circulating tumor proportions is obtained and the probability is calculated, and the significance is determined by the likelihood ratio test to determine the MRD status.

10. A micro residual lesion detection system based on ultra-high depth sequencing, characterized in that: The system is used to implement a method for detecting minimal residual lesions based on ultra-high depth sequencing as described in any one of claims 1 to 9, the system comprising: A data processing unit, used to split the sequencing library according to the sample index to obtain the sample to be tested, and obtain the consensus sequence according to the molecular identifier of the sample to be tested; A noise level evaluation unit is used to evaluate the noise level introduced into the original nucleic acid molecule chain based on the consensus sequence by filtering out the consensus sequence with a quality value less than 25 or a family size less than 3, combining the subtype type and the distance from the edge of the fragment; and evaluating the noise at the PCR level by combining the context and the chain specificity; The detection unit is used to estimate the ctDNA level based on the noise level and combined with prior knowledge of the tumor, and determine the MRD status by testing the significance of the source of the molecular signal.

Images