A method for identifying dominant clones of hbv integration sites

Abstract

The application provides a method for identifying an HBV integrated site dominant clone, comprising the following steps: obtaining and combining HBV-human integrated sites in a to-be-tested sample and a reference sample respectively to obtain a sample-level integrated site set, and calculating a clone structure index; constructing a background experience distribution by using the clone structure index of the reference sample, comparing each index of the to-be-tested sample with the background experience distribution, obtaining an experience significance measure of each clone structure index, and performing a joint significance test to obtain a joint significance measure; judging whether the to-be-tested sample has an HBV integrated dominant clone according to whether the joint significance measure meets a preset condition, and outputting a judgment result and main integrated site information. Thus, the method can accurately identify an HBV integrated dominant clone with a significant clone concentration trend, thereby reducing misjudgment caused by non-tumor background integration fluctuation, and providing technical support for screening of potential tumor-related integration events, sequencing of candidate integrated genes and subsequent function verification.

Description

Technical Field

[0001] This invention belongs to the fields of bioinformatics, genomics, and tumor genomics, specifically relating to data analysis of hepatitis B virus (HBV) integration sites and methods for identifying dominant clones associated with hepatocellular carcinoma (HCC). Background Technology

[0002] Hepatitis B virus (HBV) infection is a significant risk factor for the development of hepatocellular carcinoma (HCC). HBV DNA can integrate into the host hepatocyte genome, participating in HCC development by influencing oncogene expression, inducing genomic instability, altering chromosomal structure, or forming virus-host fusion sequences. Studies have shown that HBV integration can occur in both tumor tissues and adjacent non-tumor tissues, and some integration events are associated with oncogene activation, clonal amplification, or tumor progression.

[0003] With the development of high-throughput sequencing technology, HBV-human integration sites can be identified through whole-genome sequencing, targeted capture sequencing, transcriptome sequencing, or specialized viral integration detection procedures. Existing methods and tools, such as HIVID, Virus-Clip, and HIVID2, can locate virus-host integration breakpoints and annotate their corresponding host gene regions. Further research has revealed that HBV integration in HCC is not entirely random; some integration events can repeatedly occur in oncogenes or regulatory regions such as TERT, KMT2B / MLL4, and CCNE1. If an integration event participates in tumorigenesis or promotes tumor cell growth, cells carrying that integration event may undergo clonal expansion. This expansion, at the integration site (IS) level, can manifest as an increase in the number and proportion of reads supporting that integration site within the sample, forming a dominant clonal structure at the IS level. Simultaneously, HBV integration events can also be detected in adjacent non-tumor tissues, and a certain degree of read concentration may exist. Therefore, relying solely on the presence of a specific integration event, its location within an oncogene, or the high number of reads supported by a single major locus is insufficient to reliably distinguish between tumor-associated dominant clones and ordinary integration fluctuations in the adjacent non-tumor background. Existing HBV integration site analysis workflows struggle to objectively identify potential tumor-associated IS-level dominant clones from a large number of integration events and are prone to misinterpreting integration fluctuations in the adjacent non-tumor background as tumor-associated signals. Therefore, identifying integration events potentially related to tumor development and progression from a large pool of integration events is a pressing issue that needs to be addressed. Summary of the Invention

[0004] This invention aims to address at least one of the technical problems existing in the prior art. To this end, this invention provides a method for identifying dominant clones of HBV integration sites based on a comparison of adjacent non-tumor background distribution. This invention identifies dominant clones at the IS level that significantly deviate from the non-tumor background by using PA background distribution modeling and noise control, multi-indicator joint statistical judgment, and PA leave-one-out-of-false-positive verification. This provides a basis for screening potential tumor-related integration events, master integration sites, and candidate integration genes, and can effectively reduce misjudgments caused by integration fluctuations in the non-tumor background.

[0005] Therefore, in a first aspect, the present invention proposes a method for identifying HBV integration site dominant clones, the method comprising the following steps: S1: acquiring HBV-human integration site data of the test sample and the reference sample respectively, and merging the integration sites in each sample to obtain a sample-level integration site set; S2: calculating clonal structure indices based on the sample-level integration site set; S3: constructing a control background set using the clonal structure indices of the reference sample, and comparing the empirical distribution of each clonal structure indices of the test sample with the same clonal structure indices in the control background set to obtain empirical significance measures of each clonal structure indices; S4: performing a joint significance test on at least two of the empirical significance measures to obtain a joint significance measure; S5: determining whether the test sample has an HBV integration dominant clone based on whether the joint significance measure meets a preset significance condition, and outputting the determination result and the corresponding master integration site information; wherein, the clonal structure indices include indices characterizing the dominance of master integration site reads and indices characterizing the uneven distribution of integration reads.

[0006] In some embodiments, the sample to be tested is a tumor sample, and the reference sample is a non-tumor sample adjacent to the cancer cell.

[0007] In some embodiments, the metrics characterizing the dominance of reads at the main integration site include the proportion of reads at the main integration site and / or the cumulative proportion of reads at the top five integration sites.

[0008] In some embodiments, the indices characterizing the degree of unevenness in the distribution of integrated reads include one or more of the following: the Gini index for unevenness in the distribution of integrated reads, the uniformity index for evenness in the distribution of integrated reads, the concentration index for integrated reads, the diversity index for integrated reads, and the effective clone number.

[0009] In some embodiments, the indicators characterizing the dominance of reads at the main integration site are the proportion of reads at the main integration site and the proportion of cumulative reads at the top five integration sites, and the indicators characterizing the unevenness of the distribution of integrated reads are the Gini index of the unevenness of the distribution of integrated reads and the uniformity index of the uniformity of the distribution of integrated reads.

[0010] In some embodiments, the HBV-human integration site data includes at least the sample number, chromosome, integration location, gene annotation, and number of supporting reads.

[0011] In some embodiments, the identification method further includes the following steps: in step S1, the sample-level quality control data of the reference sample is acquired synchronously, and the reference sample is cleaned based on the quality control data; wherein, the quality control data includes at least sequencing quality, number of reads after deduplication, total number of integrated reads, and number of integration sites.

[0012] In some embodiments, the identification method further includes the following steps: in step S1, before the merging is performed, the HBV-human integration site data is initially filtered to retain only the integration site data that supports reads > a first preset threshold.

[0013] In some embodiments, the conditions for quality control cleaning include at least one of the following: (1) sequencing quality ≥ second preset threshold; (2) number of reads after deduplication ≥ preset quantile of the reference sample distribution; (3) total number of integrated reads ≥ third preset threshold; (4) number of integration sites ≥ fourth preset threshold; (5) removal of extreme single-point samples with ≤1 number of integration sites and the proportion of reads at the main integration site ≥0.95.

[0014] In some embodiments, when the number of samples to be tested is greater than 1, the identification method further includes the following step: in step S4, multiple hypothesis testing is performed on the joint significance measure to obtain the corrected significance measure.

[0015] In some embodiments, the main integration site information includes the main integration site and its corresponding host gene.

[0016] In some embodiments, the identification method further includes the following steps: querying the HBV integration database to obtain the number of times the host gene is integrated with public tumor samples and / or the number of times it is integrated with public non-tumor samples.

[0017] In some embodiments, the determination of the preset significance condition includes the following grading rules: high confidence simultaneously meets the following conditions: the joint false discovery rate is less than 0.01, the ratio of the principal site proportion to the median of adjacent non-tumor samples is greater than 2.0, and there is paired support; medium confidence meets the following conditions: the joint false discovery rate is less than 0.05, and the ratio of the principal site proportion to the median of adjacent non-tumor samples is greater than 1.

[0018] In a second aspect, the present invention proposes a system for identifying dominant HBV integration sites, the system comprising: a data acquisition module configured to acquire HBV-human integration site data of a test sample and a reference sample respectively; a site processing module configured to merge integration sites within each sample to obtain a sample-level integration site set, and calculate clonal structure indices based on the sample-level integration site set; a background comparison module configured to construct a control background set using the clonal structure indices of the reference sample, and compare the empirical distribution of each clonal structure indices of the test sample with the same clonal structure indices in the control background set to obtain empirical significance measures of each clonal structure indices; a joint test module configured to perform a joint significance test on at least two of the empirical significance measures to obtain a joint significance measure; and a judgment output module configured to determine whether the test sample has a dominant HBV integration clone based on whether the joint significance measure meets a preset significance condition, and output the judgment result and the corresponding master integration site information; wherein, the clonal structure indices include indices characterizing the dominance of master integration site reads and indices characterizing the degree of uneven distribution of integration reads.

[0019] In some embodiments, the sample to be tested is a tumor sample, and the reference sample is a non-tumor sample adjacent to the cancer cell.

[0020] In some embodiments, the metrics characterizing the dominance of reads at the main integration site include the proportion of reads at the main integration site and / or the cumulative proportion of reads at the top five integration sites.

[0021] In some embodiments, the indices characterizing the degree of unevenness in the distribution of integrated reads include one or more of the following: the Gini index for unevenness in the distribution of integrated reads, the uniformity index for evenness in the distribution of integrated reads, the concentration index for integrated reads, the diversity index for integrated reads, and the effective clone number.

[0022] In some embodiments, the indicators characterizing the dominance of reads at the main integration site are the proportion of reads at the main integration site and the proportion of cumulative reads at the top five integration sites, and the indicators characterizing the unevenness of the distribution of integrated reads are the Gini index of the unevenness of the distribution of integrated reads and the uniformity index of the uniformity of the distribution of integrated reads.

[0023] In some embodiments, the HBV-human integration site data includes at least the sample number, chromosome, integration location, gene annotation, and number of supporting reads.

[0024] In some embodiments, the site processing module is further configured to perform the following steps: acquiring sample-level quality control data of the reference sample, and performing quality control cleaning on the reference sample based on the quality control data; wherein the quality control data includes at least sequencing quality, number of deduplicated reads, total number of integrated reads, and number of integration sites.

[0025] In some embodiments, the background comparison module is further configured to perform the following steps: before the merging is performed, the HBV-human integration site data is initially filtered to retain only integration site data that supports reads > a first preset threshold.

[0026] In some embodiments, the conditions for quality control cleaning include at least one of the following: (1) sequencing quality ≥ second preset threshold; (2) number of reads after deduplication ≥ preset quantile of the reference sample distribution; (3) total number of integrated reads ≥ third preset threshold; (4) number of integration sites ≥ fourth preset threshold; (5) removal of extreme single-point samples with ≤1 number of integration sites and the proportion of reads at the main integration site ≥0.95.

[0027] In some embodiments, when the number of samples to be tested is greater than 1, the joint test module is further configured to perform the following steps: perform multiple hypothesis testing correction on the joint significance measure to obtain the corrected significance measure.

[0028] In some embodiments, the main integration site information includes the main integration site and its corresponding host gene.

[0029] In some embodiments, the system further includes an annotation module configured to query the HBV integration database to obtain the number of times the host gene is integrated with public tumor samples and / or the number of times it is integrated with public non-tumor samples.

[0030] In some embodiments, the determination of the preset significance condition includes the following grading rules: high confidence means that the following conditions are met simultaneously: the joint false discovery rate is less than 0.01, the ratio of the principal site proportion to the median of adjacent non-tumor samples is greater than 2.0, and there is paired support; medium confidence means that the following conditions are met: the joint false discovery rate is less than 0.05, and the ratio of the principal site proportion to the median of adjacent non-tumor samples is greater than 1.

[0031] In a third aspect, the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the identification method described in the first aspect.

[0032] Compared with the prior art, the beneficial technical effects achieved by the present invention are as follows: 1. Introducing adjacent non-tumor background as a reference effectively reduces false positives. This invention constructs an empirical distribution of the background of adjacent non-tumor samples (PA samples) and compares the clonal structure indicators of the tumor sample to be tested with it. This clearly distinguishes between true tumor-related dominant clones and incidental integration fluctuations in the background, significantly reducing false positives.

[0033] 2. Multi-indicator joint statistical judgment overcomes the limitations of single threshold. This invention uses multiple clone structure indicators such as top1_ratio, Gini index, and evenness index, and constructs Fisher-type joint statistics and multiple corrections to avoid instability and batch bias caused by relying on a single indicator or fixed threshold, thereby improving the robustness and objectivity of the judgment results.

[0034] 3. Provides hierarchical classification and complete interpretable output. This invention outputs four sample levels: high, medium, low confidence, and not significant. It also provides the main integration site, main integration gene, key indicator values, and public data annotations. Combined with leave-one-out-of-placing (PA) validation to assess the false positive rate, it realizes a complete chain of evidence from sample-level dominant clone determination to gene-level functional interpretation, which facilitates subsequent validation and screening.

[0035] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0036] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the description of the embodiments taken in conjunction with the following drawings, in which: Figure 1 This is a flowchart of the HBV integrated dominant clone identification method in Embodiment 1 of the present invention; Figure 2 This is a box plot of key indicators under the statistical model classification in Embodiment 2 of the present invention; Figure 3 This is a cumulative dominance curve of different categories of CA samples in Embodiment 2 of the present invention; Figure 4 This is a scatter plot of the principal site percentage and Gini index in Embodiment 2 of the present invention; Figure 5 This is a distribution diagram of the PA leave-one verification results in Embodiment 3 of the present invention. Detailed Implementation

[0037] The embodiments of the present invention are described in detail below. The embodiments described below are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0038] It should be noted that the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. Furthermore, in the description of this invention, unless otherwise stated, "a plurality of" means two or more.

[0039] The endpoints and any values ​​of the ranges disclosed herein are not limited to the precise ranges or values, and these ranges or values ​​should be understood to include values ​​close to these ranges or values. For numerical ranges, the endpoint values ​​of the various ranges, the endpoint values ​​of the various ranges and individual point values, and individual point values ​​can be combined with each other to obtain one or more new numerical ranges, which should be considered as specifically disclosed herein.

[0040] In this document, the terms “comprising” or “including” are open-ended expressions, meaning that they include the contents specified in this invention, but do not exclude other aspects.

[0041] HBV: refers to hepatitis B virus. In this invention, HBV is the viral source for host genome integration.

[0042] HCC refers to hepatocellular carcinoma. This invention is preferably applied to integrated event analysis in HBV-related hepatocellular carcinoma samples.

[0043] HBV-human integration site: refers to the location where an HBV sequence joins with a human host genome sequence. In this invention, an HBV-human integration site is typically characterized by information such as the host chromosome, integration location, gene annotation, and the number of supporting reads.

[0044] IS stands for integration site. In this invention, a dominant clone at the IS level refers to the phenomenon where a certain integration site exhibits a higher number of reads or a higher proportion of reads within the sample due to the expansion of the corresponding cell population.

[0045] PA samples: These refer to adjacent non-tumor samples, that is, samples derived from the region near tumor tissue, used to represent the integration state of the non-tumor background. In this invention, PA samples are used to construct a background reference.

[0046] CA sample: refers to the tumor sample to be tested, preferably an HBV-related hepatocellular carcinoma tissue sample. In this invention, the CA sample is the analytical object for identifying HBV integrated dominant clones.

[0047] Reads: refers to short sequence reads obtained from high-throughput sequencing. In this invention, the number of reads supporting a specific integration site is used to reflect the strength of sequencing evidence for that integration site in the sample.

[0048] top1_ratio: This refers to the proportion of reads at the master integration site to the total number of integrated reads in the sample. The higher this value, the more concentrated the reads are at a single master integration site, and the more likely there is master integration site dominance.

[0049] `top5_ratio`: This refers to the cumulative proportion of reads from the top five integration sites to the total number of integrated reads in the sample. This value reflects the contribution of head integration sites to the total integrated signal of the sample.

[0050] gini_index: This refers to the Gini index, which measures the degree of unevenness in the distribution of integrated reads across different integration sites. The higher the value, the more concentrated and uneven the distribution of integrated reads.

[0051] Evenness: This refers to the uniformity index, used to measure whether integrated reads are evenly distributed across different integration sites. The lower the value, the more concentrated the reads are at a few integration sites.

[0052] Empirical P-value: This refers to the P-value obtained by comparing a certain clonal structure index of the CA sample to be tested with the actual distribution of the corresponding index in the PA background set. In this invention, the empirical P-value does not depend on a fixed theoretical distribution, but is directly calculated based on the PA background sample.

[0053] `combined_fdr` refers to the combined false discovery rate, a statistical indicator obtained after correcting for the combined significance p-value (`combined_p`) through multiple hypothesis testing. `combined_p` is calculated by comparing multiple clonal structure indicators of the test sample (such as the predominant site percentage, Gini index, etc.) with adjacent non-tumor background samples. `combined_fdr` measures the overall deviation of the test sample from the background samples in terms of clonal structure characteristics and serves as an important grading standard for determining the presence of HBV integrated dominant clones. A smaller value indicates stronger statistical evidence of dominant clonal characteristics in the test sample, while controlling for the false discovery rate. In a first aspect of the present invention, a method for identifying HBV integration site dominant clones is proposed. The method includes the following steps: S1: acquiring HBV-human integration site data of a test sample and a reference sample, and merging the integration sites in each sample to obtain a sample-level integration site set; S2: calculating clonal structure indices based on the sample-level integration site set; S3: constructing a control background set using the clonal structure indices of the reference sample, and comparing the empirical distribution of each clonal structure indices of the test sample with the same clonal structure indices in the control background set to obtain empirical significance measures of each clonal structure indices; S4: performing a joint significance test on at least two of the empirical significance measures to obtain a joint significance measure; S5: determining whether the test sample has an HBV integration dominant clone based on whether the joint significance measure meets a preset significance condition, and outputting the determination result and the corresponding master integration site information; wherein, the clonal structure indices include indices characterizing the dominance of master integration site reads and indices characterizing the uneven distribution of integration reads.

[0054] According to an embodiment of the present invention, in step S1, HBV-human integration site data of the test sample and the reference sample are first acquired respectively. The original integration site records are initially filtered, and integration records with a number of supporting reads or a sequence count greater than a preset threshold are preferably retained. Subsequently, within a single sample, HBV-human integration site records that are the same or similar are merged according to the host chromosome, host integration breakpoint location, HBV breakpoint location and orientation information to obtain a sample-level integration site set. The number of supporting reads for each integration site, the number of integration sites in the sample, the total number of integration reads, and the read distribution of each integration site are counted. Then, in step S2, based on the sample-level integration site set, clonal structure indices used to characterize the read distribution of integration sites and their clonal advantage characteristics are calculated. Furthermore, in the presence of multiple reference samples, a control background set is constructed using the clonal structure indicators of the reference samples. In step S3, the empirical distribution of each clonal structure indicator of the test sample is compared with the corresponding indicator in the control background set to obtain the empirical significance measure of each clonal structure indicator, thereby unifying clonal structure indicators of different directions and dimensions into the degree of deviation relative to the background distribution. To reduce the one-sidedness of single indicator judgment, in step S4, at least two core clonal structure indicators' empirical significance measures are selected to construct a joint significance measure, and multiple corrections are performed on the multi-indicator comparison results to control the risk of false positives and improve the robustness of the judgment results. Finally, in step S5, the joint significance measure is combined with auxiliary indicators for multi-level classification to determine whether there is an HBV integration dominant clone in the test sample. After completing the sample-level judgment, candidate dominant integration sites and related genes are further sorted and output based on the number of supporting reads, relative abundance, degree of difference from the reference sample, and gene annotation information of each integration site, thereby forming biological conclusions that can be verified and interpreted later.

[0055] In some embodiments of the present invention, the test sample is a tumor sample, and the reference sample is an adjacent non-tumor sample. According to embodiments of the present invention, the test sample is a tumor sample (CA sample), preferably an HBV-related hepatocellular carcinoma tissue sample; the reference sample is an adjacent non-tumor sample (PA sample), derived from the region adjacent to the tumor tissue, used to represent the integration state of the non-tumor background. The PA sample of the present invention is used to construct a background reference. The PA sample originates from the adjacent tissue of the same patient and has a highly consistent genetic background, viral exposure history, and environmental exposure factors with the test tumor sample (CA), thereby effectively controlling sample heterogeneity and significantly reducing the misjudgment rate caused by fluctuations in the integration of the adjacent non-tumor background.

[0056] In some embodiments of the present invention, the indicators characterizing the dominance of reads at the master integration site include the top1 ratio of reads at the master integration site and / or the cumulative ratio of reads at the top five integration sites. According to embodiments of the present invention, the top1 ratio measures the dominance of the strongest single integration site, indicating significant expansion of the cell clone containing that site, and is the most direct indicator for identifying the "master clone." The cumulative ratio of reads at the top five integration sites (top5 ratio) measures the cumulative dominance of the top few sites, capable of capturing cases of "oligoclones" or "multiple clones with highly concentrated head sites," compensating for the shortcomings of a single top1 ratio. Therefore, using these two indicators in combination provides complementary information dimensions for subsequent joint statistics, improving the sensitivity and robustness of determining the dominant clone.

[0057] In some embodiments of the present invention, the indices characterizing the degree of unevenness in the distribution of integrated reads include one or more of the following: the Gini index for unevenness in the distribution of integrated reads, the uniformity index for evenness in the distribution of integrated reads, the concentration index for integrated reads, the diversity index for integrated reads, and the effective clone number.

[0058] According to embodiments of the present invention, the Gini index is used to reflect the degree of uneven distribution of integrated reads among different integration sites. A higher Gini index indicates that reads are more concentrated in a few sites. The evenness index is used to measure the evenness of the distribution of integrated reads among various integration sites. A lower evenness index indicates that reads are more concentrated in a few sites, suggesting the possible existence of dominant clones. The concentration index (HHI index) is used to characterize the concentration of integrated reads. A higher HHI index indicates a more obvious dominant site. The diversity index (Shannon index) is used to characterize the diversity of the distribution of integration sites. A lower Shannon index indicates a more concentrated clonal distribution. The effective clone number is used to characterize the number of clones with equivalent contributions in the sample. A smaller value indicates that a few integration sites dominate. The above indicators can be used as primary or auxiliary indicators to jointly characterize the distribution characteristics of integrated reads from different statistical dimensions, thereby improving the consistency, robustness, and interpretability of the dominant clone identification results, and providing support for subsequent result interpretation and experimental verification.

[0059] In some embodiments of the present invention, the indicators characterizing the dominance of reads at the main integration site are the proportion of reads at the main integration site and the proportion of cumulative reads at the top five integration sites, and the indicators characterizing the unevenness of the distribution of integrated reads are the Gini index of the unevenness of the distribution of integrated reads and the uniformity index of the uniformity of the distribution of integrated reads.

[0060] In some embodiments of the present invention, the HBV-human integration site data includes at least the sample number, chromosome, integration location, gene annotation, and number of supporting reads.

[0061] In some embodiments of the present invention, the identification method further includes the following steps: in step S1, sample-level quality control data of the reference sample is acquired synchronously, and the reference sample is cleaned based on the quality control data; wherein, the quality control data includes at least sequencing quality, number of deduplicated reads, total number of integrated reads, and number of integration sites. According to embodiments of the present invention, the present invention synchronously acquires quality control data of the reference sample and performs quality control cleaning to remove low-quality, weakly signaled, or extremely abnormal PA samples, ensuring that the PA background set can truly and stably represent the HBV integration background state in non-tumor tissues, thereby avoiding misjudgment or missed judgment of dominant clones in tumor samples due to background set contamination. In the present invention, the cleaned PA samples constitute the PA background set for subsequent empirical distribution comparison and false positive verification. Therefore, the method of the present invention has high specificity.

[0062] In some embodiments of the present invention, the identification method further includes the following step: in step S1, before the merging is performed, the HBV-human integration site data is initially filtered, retaining only integration site data with a support read count > a first preset threshold. According to embodiments of the present invention, by performing initial filtering to avoid interference with subsequent intrasample merging, clonal structure index calculation, and background empirical distribution construction, the accuracy of identifying truly dominant clones is improved.

[0063] In some embodiments of the present invention, the conditions for quality control cleaning include at least one of the following: (1) sequencing quality ≥ a second preset threshold; (2) the number of reads after deduplication ≥ a preset quantile of the reference sample distribution; (3) the total number of integrated reads ≥ a third preset threshold; (4) the number of integration sites ≥ a fourth preset threshold; (5) removing extreme single-point samples with ≤1 integration site and a major integration site read ratio ≥ 0.95. This ensures that the constructed background reference set (PA background set) can realistically and stably represent the HBV integration background status in non-tumor tissues, thereby providing a reliable benchmark for subsequent empirical distribution comparison and dominant clone determination.

[0064] In some embodiments of the present invention, the identification method further includes: in step S3, the empirical distribution comparison includes right-tailed empirical P-values ​​and left-tailed empirical P-values. According to embodiments of the present invention, for high-ranking indicators, including at least one of top1_ratio, top5_ratio, Gini index, and concentration index (hhi_index), the right-tailed empirical P-value is used:

[0065] For low-tailed indicators, including at least one of evenness, shannon index, and effective clone number, the left-tailed empirical p-value is used:

[0066] Where m is the number of samples in the PA background set, M CA,j For the CA sample to be tested j Each indicator value, M PA,i,j For the first i The corresponding PA sample of the first j Individual indicator values.

[0067] Through step S3, this invention can uniformly convert clonal structure indices of different directions and dimensions into the degree of deviation relative to the PA background distribution.

[0068] In some embodiments of the present invention, the identification method further includes: in step S4, the method for obtaining the joint significance measure includes: selecting empirical significance measures (empirical P-values) of at least two core clonal structure indices to construct a Fisher-type joint statistic, preferably, the core clonal structure indices include at least one of top1_ratio, top5_ratio, Gini_index and evenness, and calculating the joint significance measure (combined_p) according to the theoretical distribution of the Fisher-type joint statistic; or, obtaining the empirical zero distribution of the joint statistic in the PA background through leave-one-out verification or background permutation, and calculating the empirically calibrated joint significance measure accordingly.

[0069] This invention uses the following formula to construct the joint statistic:

[0070] in, p j For the first j The empirical p-value of each core indicator k The number of core indicators.

[0071] In some embodiments of the present invention, when the number of samples to be tested is greater than 1, the identification method further includes the following step: in step S4, multiple hypothesis testing is performed on the joint significance measure to obtain the corrected significance measure. According to an embodiment of the present invention, Benjamini-Hochberg multiple correction is performed on the joint significance measure (combined_p) of multiple CA samples to be tested to obtain combined_fdr, which is used to measure the comprehensive significance deviation of the sample to be tested relative to the background distribution of PA.

[0072] In some embodiments of the present invention, the identification method further includes: dominant clonal classification and main integration site output. In this invention, based on combined_fdr, clonal concentration index, and pairing support information, the CA sample to be tested is classified into dominant clonals. The classification results include at least one of high confidence, medium confidence, low confidence review, and not significant. While outputting the sample-level classification results, this invention further outputs corresponding main integration site, main integration gene, main site reads percentage, top1_ratio, Gini index, combined_p, and combined_fdr, thereby achieving a correspondence from sample-level dominant clonal determination to site-level and gene-level interpretation.

[0073] In some embodiments of the present invention, the master integration site information includes the master integration site and its corresponding host gene. Therefore, the present invention can further identify HBV integration-dominant clones with a significant clonal concentration trend based on existing integration site detection results, thereby reducing misjudgments caused by integration fluctuations in non-tumor backgrounds.

[0074] In some embodiments of the present invention, the identification method further includes the following steps: querying an HBV integration database to obtain the number of integration support counts for the host gene in public tumor samples and / or public non-tumor samples. According to embodiments of the present invention, after obtaining the dominant clone classification results of the CA sample to be tested, the main integrating gene can be externally annotated by combining publicly available HBV tumor and non-tumor integration data. This public data annotation is used to provide tumor support, non-tumor support, and candidate gene interpretation information. Through this step, recurring candidate integrating genes in tumor samples and common duplicate integration events in non-tumor backgrounds can be further distinguished, providing a reference for subsequent functional validation and candidate target gene sequencing.

[0075] In some embodiments of the present invention, the determination of the preset significance condition includes the following hierarchical determination rules: According to embodiments of the present invention, high confidence (high_confidence) satisfies the following conditions: combined false discovery rate (combined_fdr) is less than 0.01, the ratio of the principal site percentage to the median of adjacent non-tumor samples is greater than 2.0 (top1_ratio_vs_pa_median_ratio>2.0), and paired support (paired_support is true); medium confidence (medium_confidence) satisfies the following conditions: combined false discovery rate is less than 0.05, and the ratio of the principal site percentage to the median of adjacent non-tumor samples is greater than 1. The grading rules of the present invention, by setting grading thresholds from "strict" to "lenient," distinguish between "core-driven clones" under high confidence and "generally suspicious clones" under medium confidence, thereby providing a clear and operable basis for subsequent scientific research and clinical decision-making. In a second aspect of the present invention, a system for identifying dominant HBV integration sites is proposed. The system includes: a data acquisition module configured to acquire HBV-human integration site data of a test sample and a reference sample, respectively; a site processing module configured to merge integration sites within each sample to obtain a sample-level integration site set, and calculate clonal structure indices based on the sample-level integration site set; a background comparison module configured to construct a control background set using the clonal structure indices of the reference sample, and compare the empirical distribution of each clonal structure indices of the test sample with the same clonal structure indices in the control background set to obtain empirical significance measures of each clonal structure indices; a joint test module configured to perform a joint significance test on at least two of the empirical significance measures to obtain a joint significance measure; and a judgment output module configured to determine whether the test sample has a dominant HBV integration clone based on whether the joint significance measure meets a preset significance condition, and output the judgment result and the corresponding master integration site information; wherein the clonal structure indices include indices characterizing the dominance of master integration site reads and indices characterizing the uneven distribution of integration reads.

[0076] Those skilled in the art will understand that the features and advantages described above for the identification method of HBV integration site dominant clones are also applicable to this system, and will not be repeated here.

[0077] In some embodiments of the present invention, the sample to be tested is a tumor sample, and the reference sample is a non-tumor sample adjacent to the cancer cell.

[0078] In some embodiments of the present invention, the indicators characterizing the dominance of reads at the main integration site include the proportion of reads at the main integration site and / or the cumulative proportion of reads at the top five integration sites.

[0079] In some embodiments of the present invention, the indices characterizing the degree of unevenness in the distribution of integrated reads include one or more of the following: the Gini index for unevenness in the distribution of integrated reads, the uniformity index for evenness in the distribution of integrated reads, the concentration index for integrated reads, the diversity index for integrated reads, and the effective clone number.

[0080] In some embodiments of the present invention, the indicators characterizing the dominance of reads at the main integration site are the proportion of reads at the main integration site and the proportion of cumulative reads at the top five integration sites, and the indicators characterizing the unevenness of the distribution of integrated reads are the Gini index of the unevenness of the distribution of integrated reads and the uniformity index of the uniformity of the distribution of integrated reads.

[0081] In some embodiments of the present invention, the HBV-human integration site data includes at least the sample number, chromosome, integration location, gene annotation, and number of supporting reads.

[0082] In some embodiments of the present invention, the site processing module is further configured to perform the following steps: obtaining sample-level quality control data of the reference sample, and performing quality control cleaning on the reference sample based on the quality control data; wherein the quality control data includes at least sequencing quality, number of deduplicated reads, total number of integrated reads, and number of integration sites.

[0083] In some embodiments of the present invention, the background comparison module is further configured to perform the following steps: before the merging is performed, the HBV-human integration site data is initially filtered to retain only the integration site data that supports reads > a first preset threshold.

[0084] In some embodiments of the present invention, the conditions for quality control cleaning include at least one of the following: (1) sequencing quality ≥ second preset threshold; (2) number of reads after deduplication ≥ preset quantile of the reference sample distribution; (3) total number of integrated reads ≥ third preset threshold; (4) number of integration sites ≥ fourth preset threshold; (5) removal of extreme single-point samples with ≤1 number of integration sites and the proportion of reads at the main integration site ≥0.95.

[0085] In some embodiments of the present invention, when the number of samples to be tested is greater than 1, the joint test module is further configured to perform the following steps: perform multiple hypothesis testing correction on the joint significance measure to obtain the corrected significance measure.

[0086] In some embodiments of the present invention, the main integration site information includes the main integration site and its corresponding host gene.

[0087] In some embodiments of the present invention, the system further includes an annotation module configured to query the HBV integration database to obtain the number of times the host gene is integrated with public tumor samples and / or the number of times it is integrated with public non-tumor samples.

[0088] In some embodiments of the present invention, the determination of the preset significance condition includes the following hierarchical determination rules: High confidence means that the following conditions are met simultaneously: the joint false discovery rate is less than 0.01, the ratio of the proportion of the principal site to the median of adjacent non-tumor samples is greater than 2.0, and there is paired support; medium confidence means that the following conditions are met: the joint false discovery rate is less than 0.05, and the ratio of the proportion of the principal site to the median of adjacent non-tumor samples is greater than 1.

[0089] In a third aspect of the present invention, the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the identification method described in the first aspect.

[0090] Those skilled in the art will understand that the features and advantages described above for the identification method and system for HBV integration site dominant clones are also applicable to this computer-readable storage medium, and will not be repeated here.

[0091] The present invention will be explained below with reference to embodiments. Those skilled in the art will understand that the following embodiments are for illustrative purposes only and should not be considered as limiting the scope of the invention. Where specific techniques or conditions are not specified in the embodiments, they are performed according to the techniques or conditions described in the literature in the field or according to the product instructions. Reagents or instruments whose manufacturers are not specified are all conventional products that can be obtained commercially.

[0092] Example 1: Raw integrated data processing, intra-sample merging, and PA background set construction This embodiment aims to illustrate how the present invention converts HBV-human integration site detection results into sample-level standardized data that can be used for dominant clone identification, and constructs a PA background set for subsequent empirical distribution comparison. Combined with... Figure 1 The process shown in this embodiment includes the following steps.

[0093] 1.1 Data Input and Preliminary Filtering Obtain the standardized results table for HBV-human integration sites and the sample-level quality control index table. Integration site data includes sample number, sample type, chromosome, integration location, gene annotation, and number of supporting reads; sample-level quality control data includes sequencing quality, number of deduplicated reads, number of integration sites, and total number of integrated reads.

[0094] The original integration site records are initially filtered. Preferably, only integration records with a Sequence Count greater than 5 are retained to reduce the impact of low-support records, sequencing noise, or false positive alignments on subsequent analyses.

[0095] 1.2 Intrasample integration site merging Within each sample, records of identical or similar integration sites are merged according to chromosome, integration location, and gene annotation to obtain a sample-level set of integration sites. For each sample, information such as the number of integration sites, total number of integration reads, number of reads at major integration sites, and the proportion of major sites are statistically analyzed for subsequent calculations of clonal structure indicators.

[0096] 1.3 PA Background Cleaning and Background Set Construction After completing intrasample merging and preliminary index calculation, background cleaning of PA samples is performed based on sample-level quality control data and clone structure indicators. Preferred cleaning conditions include: raw_q30_ratio_pct not less than 80, after_rmdup_trim_count not less than the 4th percentile of the PA sample distribution, total_reads not less than 20, sites not less than 3, integration_total_sequence_counts not less than 20, and extreme single-point samples with sites less than or equal to 1 and top1_ratio greater than or equal to 0.95 are removed.

[0097] After the above processing, the original number of PA samples in this embodiment was 106, and 91 samples were retained as the PA background set after cleaning. The number of CA samples was 107, and the number of CA-PA paired samples was 84. Table 1 summarizes the background cleaning and sample overview in this embodiment.

[0098] Table 1. PA Background Cleaning and Sample Overview ; As shown in Table 1, the median top1_ratio of the CA samples is higher than that of the PA background samples, while the median evenness of the CA samples is lower than that of the PA background samples, suggesting that the CA samples are more likely to show a concentration of major integration sites. The 91 PA background cases constructed in this embodiment are used for subsequent empirical distribution comparisons, joint statistic calculations, and leave-one-out validation of PA.

[0099] Example 2: The ability of clonal structure indicators to distinguish CA / PA and the classification of dominant clones. This embodiment aims to illustrate the clonal structure index and multi-index joint statistical method proposed in this invention, reflecting the differences in integrated clonal structure between CA samples and PA background samples at the IS level, and classifying the dominant clones of the CA samples to be tested based on this. Figures 2 to 4This embodiment includes the following steps.

[0100] 2.1 Calculation of Clonal Structure Indicators and Assessment of CA / PA Differences Based on the sample-level integration site set obtained in Example 1, clonal structure indices were calculated for each sample. The core indices include top1_ratio, top5_ratio, gini_index, and evenness; where top1_ratio reflects the dominance of the major integration site, top5_ratio reflects the cumulative contribution of the head integration site, gini_index reflects the unevenness of the integration read distribution, and evenness reflects the uniformity of the integration read distribution.

[0101] This embodiment first compares the overall distribution differences of the above indicators between CA samples and PA background samples. Generally, if an HBV integration event is accompanied by tumor cell proliferation, the integration event may manifest at the IS level as an increased proportion of main integration site reads, an increased proportion of head integration sites, an increased Gini index, and a decreased evenness. Therefore, if the CA samples show a combination of the above indicators that deviates significantly from the PA background, it can serve as a clue for screening potential tumor-related integration events and dominant clones at the IS level.

[0102] like Figure 2 As shown in the box plots of key indicators under the statistical model classification, there are differences in the distribution of top1_ratio, top5_ratio, gini_index, and evenness among different categories of samples. High-confidence and medium-confidence CA samples typically show higher top1_ratio, top5_ratio, and gini_index, as well as lower evenness, suggesting that their integrated reads are more concentrated on one or a few IS sites; while the non-significant samples are generally closer to the PA background distribution.

[0103] 2.2 Calculation of Single-Indicator P-value Based on PA Background Empirical Distribution To avoid batch bias caused by relying solely on fixed thresholds, this embodiment compares the clonal structure index of each CA sample to be tested with the empirical distribution of the PA background set. Let the PA background set contain m samples, and let the value of the j-th index in the CA sample to be tested be denoted as M_CA,j. i The corresponding value in each PA sample is denoted as M. PA,i,j For high-tail metrics such as top1_ratio, top5_ratio, gini_index, and hhi_index, the empirical P-value for the right tail is defined as:

[0104] For low-tailed metrics such as evenness, shannon index, and effective clone number, the left-tailed empirical p-value is defined as:

[0105] +1 and m+1 are used to avoid the empirical P value being 0 and to ensure the stability of subsequent joint statistics calculations.

[0106] 2.3 Fisher-type joint statistics and multiple corrections To comprehensively evaluate the contribution of multiple clonal structure indicators to the CA / PA distinction, empirical P-values ​​p_1,...,p_k of k core indicators were selected to construct a Fisher-type joint statistic.

[0107] A larger T-value indicates a higher degree of comprehensive deviation of the test sample from the PA background. In one implementation, the joint significance P-value (combined_p) is obtained based on the chi-square approximation; in a preferred implementation, the empirical distribution of T under the PA background is obtained through leave-one-out PA or background permutation, and the empirically calibrated joint significance P-value is calculated accordingly. Subsequently, the joint significance P-values ​​of multiple test samples are subjected to Benjamini-Hochberg correction to obtain combined_fdr.

[0108] The classification uses a three-tier rule: high confidence (high_confidence) satisfies combined_fdr < 0.01, top1_ratio_vs_pa_median_ratio > 2.0, and paired_support is true; medium confidence (medium_confidence) satisfies combined_fdr < 0.05, top1_ratio_vs_pa_median_ratio > 1.5; low confidence (review) satisfies combined_p < 0.05; and the remaining samples are denoted as not_significant.

[0109] 2.4 CA Sample Classification Results and Algorithm Discrimination Performance In this embodiment, after the 107 CA samples were judged by the method of this invention, 22 cases had high confidence, 30 cases had medium confidence, 16 cases had low confidence, and 39 cases were not significant. Table 2 summarizes the formal classification results of CA and the leave-one-out validation summary of PA. The leave-one-out validation results of PA show that there was 1 high-confidence false positive and 7 medium-confidence or higher false positives in the PA background samples, indicating that the method can simultaneously evaluate the false positives of PA background when identifying IS-level dominant clones that significantly deviate from the PA background in CA.

[0110] Table 2. Statistics on CA / PA Classification and Leave-one-out Verification ; Table 3 lists representative high-confidence CA samples. As shown in Table 3, the median top1_ratio, gini_index, and top1_ratio / PA of these representative high-confidence samples are all high, combined_p and combined_fdr reach the formal classification threshold, and all have pairing support information. This indicates that the main integration site accounts for a high proportion of the samples, and its multi-indicator combination shows a significant deviation relative to the PA background.

[0111] Table 3 Representative CA Samples ; like Figure 3 As shown, the cumulative dominance curves for different categories of CA samples illustrate the changes in the cumulative proportion of head integration sites. The curves for high-confidence samples typically approach 1 more quickly, indicating that a few head IS sites have already occupied the majority of integrated reads; the curves for non-significant samples are relatively flat, suggesting that their integration distribution is more dispersed.

[0112] like Figure 4 As shown, the scatter plot of principal site proportion and Gini index shows the concentration differences of various types of samples from a two-dimensional perspective. High confidence samples are usually located in regions where both top1_ratio and gini_index are high, while non-significant samples are more distributed in regions with low top1_ratio or low gini_index.

[0113] Therefore, this embodiment shows that the clonal structure index and multi-index joint statistical method proposed in this invention can effectively reflect the difference in clonal concentration at the IS level between CA samples and PA background samples, and can be used to identify HBV integration dominant clones that deviate significantly from the non-tumor background in CA samples, providing a basis for screening potential tumor-related integration events, main integration sites and candidate integration genes.

[0114] Example 3: Model accuracy assessment, leave-one-out validation of PA, and annotation of common data This embodiment aims to illustrate that the model obtained by comparing the empirical distribution of PA background and using multi-indicator joint statistical judgment can achieve good differentiation between CA samples and PA background samples. The model's misclassification level in non-tumor backgrounds is evaluated through PA leave-one-out validation. Furthermore, this embodiment also incorporates external annotation of candidate master integration genes using public HBV integrated data to improve the completeness of result interpretation. Figure 5 This embodiment includes the following steps.

[0115] 3.1 Model Performance Evaluation Based on CA / PA Discrimination Results In Example 2, the present invention performed dominant clone classification on 107 CA samples, obtaining 22 high-confidence samples, 30 medium-confidence samples, 16 low-confidence retest samples, and 39 unsignificant samples. This result demonstrates that the present invention can identify a subset of IS-level dominant clone samples from CA samples that significantly deviate from the PA background.

[0116] To evaluate the model's ability to distinguish non-tumor backgrounds, this embodiment further employs leave-one-out validation (PA). Specifically, one PA sample is removed from the PA background set each time, and this sample is re-entered as the test sample for empirical distribution comparison, Fisher-type joint statistic calculation, and formal classification steps. The remaining PA samples are treated as the background set. By counting the number of PA samples that are misclassified as high-confidence or medium-confidence or higher, the model's false positive level and specificity in non-tumor backgrounds are evaluated.

[0117] In this embodiment, leave-one-out validation results for 91 PA background samples showed 1 high-confidence classification false positive and 7 medium-confidence and above classification false positives. Therefore, at the high-confidence level, the PA background specificity is 90 / 91, approximately 98.9%; at the medium-confidence and above level, the PA background specificity is 84 / 91, approximately 92.3%. These results demonstrate that the model of this invention has a low false positive rate in non-tumor backgrounds and can effectively distinguish between dominant clonal signals significantly deviating from the PA background in CA samples and ordinary integration fluctuations in the PA background.

[0118] like Figure 5 As shown in the figure, the PA leave-one-out validation result distribution plot illustrates the composition of the classification results of PA samples in the leave-one-out test. The plot visually demonstrates that most PA samples fall into the non-significant category, with only a small number entering the high-confidence or medium-confidence categories or higher. This indicates that the model of this invention can effectively control the risk of misclassification in non-tumor background samples.

[0119] 3.2 Interpretation of High-Confidence Samples and Master Integration Sites For CA samples determined to have high or medium confidence, this invention further outputs information such as the main integration site, the main integration gene, the proportion of reads at the main site, top1_ratio, gini_index, combined_p, combined_fdr, the median of top1_ratio / PA, and paired_support. Thus, this invention enables a correspondence between sample-level dominant clone determination and site-level and gene-level interpretation.

[0120] The HBV integration dominant clone referred to in this invention is an integration clone structure in which one or a few HBV integration sites occupy the majority of the number of major integration reads in a single sample, and its clonal structure indices deviate significantly from the empirical distribution of the PA background. This dominant clone is further characterized by the major integration site, the major integration gene, the proportion of major site reads, and joint significance indices.

[0121] 3.3 External Annotations for Public Data After obtaining the official classification results of the CA samples, this invention also combines publicly available integrated HBV tumor and non-tumor data to perform external annotation on the main integrated genes. Public data annotation is used to provide information on the number of times public tumor support is received, the number of times public non-tumor support is received, and the interpretation of the results, but it is not a prerequisite for the main statistical determination. Table 4 lists the public data support for representative genes.

[0122] Table 4 External Validation of Public Data ; As shown in Table 4, TERT, KMT2B, and CCNE1 received more support in public tumor samples than in public non-tumor samples, and can be used as evidence of tumor enrichment consistency. FN1 received a relatively high number of support in public non-tumor samples, therefore it should be handled with caution in the interpretation of results. Through this step, the present invention can combine sample-level dominant clone determination, site-level master integration events, and gene-level public data support, thereby improving the completeness of result interpretation.

[0123] This embodiment demonstrates that the model established in this invention exhibits good overall performance in distinguishing between CA and PA: at medium confidence levels and above, it can identify 52 CA candidate samples while maintaining approximately 92.3% PA background specificity, and at high confidence levels, the PA background specificity reaches approximately 98.9%. This indicates that the model of this invention can identify IS-level dominant clones in CA samples that significantly deviate from the PA background while controlling for misclassification in non-tumor backgrounds, providing support for screening potential tumor-related integrative events and interpreting candidate genes.

[0124] In summary, the method of this invention no longer relies solely on the number of reads at a single master integration site, a fixed threshold, or manual experience. Instead, it uses adjacent non-tumor samples (PA samples) as a non-tumor background reference and performs sample-level clonal structure quantification and background distribution comparison on the tumor samples to be tested (CA samples). By filtering HBV-human integration site data, merging integration sites within samples, calculating clonal structure indicators, calculating empirical P-values, constructing Fisher-type joint statistics, performing multiple corrections, and verifying PA with one false positive, this invention can further identify HBV integration-dominant clones with significant clonal concentration trends based on existing integration site detection results, reducing misjudgments caused by integration fluctuations in the non-tumor background, thereby providing a basis for screening potential tumor-related integration events, master integration sites, and candidate integration genes.

[0125] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0126] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.

Claims

1. A method for identifying dominant clones at HBV integration sites, characterized in that, The identification method includes the following steps: S1: Obtain HBV-human integration site data for the test sample and the reference sample respectively, and merge the integration sites in each sample to obtain a sample-level integration site set. S2: Calculate the clonal structure index based on the sample-level integration site set; S3: Construct a control background set using the clonal structure indices of the reference sample, and compare the empirical distribution of each clonal structure index of the test sample with the same clonal structure index in the control background set to obtain the empirical significance measure of each clonal structure index. S4: Perform a joint significance test on at least two of the empirical significance measures to obtain a joint significance measure; S5: Based on whether the joint significance measure meets the preset significance condition, determine whether the sample to be tested has an HBV integration dominant clone, and output the judgment result and the corresponding main integration site information; The clonal structure indices include indices characterizing the dominance of reads at the major integration site and indices characterizing the uneven distribution of integrated reads.

2. The identification method according to claim 1, characterized in that, The sample to be tested is a tumor sample, and the reference sample is a non-tumor sample adjacent to the cancer cell.

3. The identification method according to claim 1, characterized in that, The indicators characterizing the dominance of reads at the main integration site include the proportion of reads at the main integration site and / or the cumulative proportion of reads at the top five integration sites. Optionally, the index characterizing the degree of unevenness in the distribution of integrated reads includes one or more of the following: the Gini index of unevenness in the distribution of integrated reads, the evenness index of the distribution of evenness in integrated reads, the concentration index of integrated reads, the diversity index of integrated reads, and the effective clone number. Optionally, the indicators characterizing the dominance of reads at the main integration site are the proportion of reads at the main integration site and the proportion of cumulative reads at the top five integration sites, and the indicators characterizing the unevenness of the distribution of integrated reads are the Gini index of the unevenness of the distribution of integrated reads and the uniformity index of the uniformity of the distribution of integrated reads.

4. The identification method according to claim 1, characterized in that, The HBV-human integration site data shall include at least the sample number, chromosome, integration location, gene annotation, and number of supporting reads; Optionally, the identification method further includes the following steps: In step S1, the sample-level quality control data of the reference sample is acquired synchronously, and the reference sample is cleaned based on the quality control data. The quality control data includes at least sequencing quality, number of deduplicated reads, total number of integrated reads, and number of integration sites. Optionally, the identification method further includes the following steps: In step S1, before the merging is performed, the HBV-human integration site data is initially filtered, and only integration site data with a number of supported reads > a first preset threshold is retained; Optionally, the conditions for the quality control cleaning include at least one of the following: (1) Sequencing quality ≥ second preset threshold; (2) The number of reads after deduplication is greater than or equal to the preset quantile of the reference sample distribution; (3) The total number of integrated reads is greater than or equal to the third preset threshold; (4) The number of integration sites is greater than or equal to the fourth preset threshold; (5) Remove extreme single-point samples with ≤1 number of integration sites and ≥0.95% of reads from major integration sites.

5. The identification method according to claim 1, characterized in that, When the number of samples to be tested is greater than 1, the identification method further includes the following steps: In step S4, the joint significance measure is corrected by multiple hypothesis testing to obtain the corrected significance measure.

6. The identification method according to claim 1, characterized in that, The main integration site information includes the main integration site and its corresponding host gene; Optionally, the identification method further includes the following steps: Query the HBV integration database to obtain the number of times the host gene is supported by integration of public tumor samples and / or public non-tumor samples.

7. The identification method according to claims 1 to 6, characterized in that, The determination of the preset significance condition includes the following hierarchical determination rules: High confidence is achieved when the following conditions are met: the joint false discovery rate is less than 0.01, the ratio of the proportion of principal sites to the median of adjacent non-tumor samples is greater than 2.0, and there is paired support. The medium confidence level meets the following conditions: the joint false discovery rate is less than 0.05, and the ratio of the proportion of principal sites to the median of adjacent non-tumor samples is greater than 1.

8. A system for identifying dominant clonal variants at HBV integration sites, characterized in that, The system includes: The data acquisition module is configured to acquire HBV-human integration site data of the test sample and the reference sample, respectively. The site processing module is configured to merge integration sites within each sample to obtain a sample-level integration site set, and to calculate clonal structure indices based on the sample-level integration site set. The background comparison module is configured to construct a control background set using the clonal structure indices of the reference sample, and to compare the empirical distribution of each clonal structure index of the test sample with the same clonal structure index in the control background set to obtain the empirical significance measure of each clonal structure index. The joint test module is configured to perform a joint significance test on at least two of the empirical significance measures to obtain a joint significance measure; The determination output module is configured to determine whether the sample to be tested has an HBV integration advantage clone based on whether the joint significance measure meets the preset significance condition, and output the determination result and the corresponding main integration site information. The clonal structure indices include indices characterizing the dominance of reads at the major integration site and indices characterizing the uneven distribution of integrated reads.

9. The system according to claim 8, characterized in that, The sample to be tested is a tumor sample, and the reference sample is a non-tumor sample adjacent to the cancer cell. Optionally, the indicators characterizing the dominance of reads at the main integration site include the proportion of reads at the main integration site and / or the cumulative proportion of reads at the top five integration sites. Optionally, the index characterizing the degree of unevenness in the distribution of integrated reads includes one or more of the following: the Gini index of unevenness in the distribution of integrated reads, the evenness index of the distribution of evenness in integrated reads, the concentration index of integrated reads, the diversity index of integrated reads, and the effective clone number. Optionally, the indicators characterizing the dominance of reads at the main integration site are the proportion of reads at the main integration site and the cumulative proportion of reads at the top five integration sites, and the indicators characterizing the unevenness of the distribution of integrated reads are the Gini index of the unevenness of the distribution of integrated reads and the uniformity index of the uniformity of the distribution of integrated reads. Optionally, the HBV-human integration site data includes at least the sample number, chromosome, integration location, gene annotation, and number of supporting reads; Optionally, the site processing module is further configured to perform the following steps: Obtain the sample-level quality control data of the reference sample, and perform quality control cleaning on the reference sample based on the quality control data; The quality control data includes at least sequencing quality, number of deduplicated reads, total number of integrated reads, and number of integration sites. Optionally, the background comparison module is further configured to perform the following steps: Before the merging process, the HBV-human integration site data is initially filtered, retaining only integration site data with a number of supported reads greater than a first preset threshold. Optionally, the conditions for the quality control cleaning include at least one of the following: (1) Sequencing quality ≥ second preset threshold; (2) The number of reads after deduplication is greater than or equal to the preset quantile of the reference sample distribution; (3) The total number of integrated reads is greater than or equal to the third preset threshold; (4) The number of integration sites is greater than or equal to the fourth preset threshold; (5) Remove extreme single-point samples with ≤1 integration site and ≥0.95% of reads from major integration sites; Optionally, when the number of samples to be tested is greater than 1, the joint testing module is further configured to perform the following steps: Multiple hypothesis testing is performed on the joint significance measure to obtain the corrected significance measure; Optionally, the main integration site information includes the main integration site and its corresponding host gene; Optionally, the system further includes an annotation module configured to query the HBV integration database to obtain the number of times the host gene is integrated with public tumor samples and / or the number of times it is integrated with public non-tumor samples. Optionally, the determination of the preset saliency condition includes the following hierarchical determination rules: High confidence requires meeting the following conditions: a combined false discovery rate of less than 0.01, a ratio of the proportion of principal sites to the median of adjacent non-tumor samples greater than 2.0, and paired support; medium confidence requires meeting the following conditions: a combined false discovery rate of less than 0.05, and a ratio of the proportion of principal sites to the median of adjacent non-tumor samples greater than 1.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, which, when executed by a processor, implements the identification method according to any one of claims 1 to 7.

Images