A method for detecting integration sites and quantitatively detecting viral integration efficiency based on targeted capture sequencing

Abstract

This invention provides a method for detecting integration sites and quantitatively assessing viral integration efficiency based on targeted capture sequencing. The method includes targeted enrichment and sequencing of samples using a targeted capture sequencing panel; performing virus-host dual-reference alignment and breakpoint clustering on the enriched sequencing data to achieve comprehensive detection of integration sites; and calculating the integration efficiency by comparing the abundance of virus-host fusion reads at each integration site with the total amount of virus-specific reads. This allows for simultaneous quantitative assessment of integration site profiles and integration efficiency in the same sequencing experiment. Compared to the LAM-PCR method, the TES Panel of this invention significantly improves the sensitivity and comprehensiveness of fragmented integration detection through multi-site, multi-fragment targeted capture, providing key technical support for quality control of gene therapy products, risk warning of viral infection-related diseases, and safety assessment.

Description

Technical Field

[0001] This invention belongs to the field of biomedical technology, specifically, it relates to a method for detecting integration sites and quantitatively detecting viral integration efficiency based on targeted capture sequencing. Background Technology

[0002] With the rapid development of gene therapy and viral vector technologies, the introduction of exogenous genes into host cells using vectors such as adeno-associated virus (AAV) and lentiviruses has become an important direction for the treatment of various genetic diseases and tumors. During gene editing or virus-mediated transduction, the insertion of exogenous DNA into the genome is usually accomplished through the host cell's own intracellular repair mechanisms. Numerous studies have shown that viral genomes can undergo random or selective integration events into host DNA, with some insertion sites located near oncogenes or regulatory elements, which are associated with clonal amplification and the risk of tumorigenesis. Therefore, systematic monitoring of viral integration sites is necessary.

[0003] Currently, viral integration detection mainly relies on two strategies: one is based on PCR, such as linear amplification-mediated PCR (LAM-PCR), modified nrLAM-PCR, and ligation-mediated PCR. These methods amplify unknown host-side sequences using primers on known viral sequences, and then combine this with high-throughput sequencing to resolve integration sites. However, due to their reliance on specific primers and restriction endonuclease sites, these methods are susceptible to factors such as the distribution of restriction sites, fragment length, and viral structural deletions, resulting in incomplete detection of integration sites and a high false-negative rate. The other approach uses whole-genome sequencing (WGS) to search for host-virus chimeric reads across the entire genome, thereby obtaining a more unbiased profile of integration sites. However, in clinical samples, the frequency of integration events is usually low, and without significantly increasing sequencing depth, WGS still has significant limitations in terms of detection sensitivity and cost.

[0004] Recent studies on hepatitis B virus (HBV) and AAV vectors have revealed that the virus often exists in complex "fragmented integration" forms, such as multiple fragments, inversions, deletions, or host-virus-host integration. Single primer amplification is unlikely to cover all breakpoint structures. In these cases, traditional methods not only easily miss integration sites but also significantly underestimate the true integration load in the sample, thus affecting the objective assessment of gene therapy products and the risk of HBV-related tumors.

[0005] To overcome the aforementioned limitations, some studies have attempted to use targeted enrichment sequencing (TES) to enrich virus-related sequences, thereby improving the sensitivity and specificity of integration detection. However, existing targeting methods mostly focus on whether the integration site is detected, and a systematic solution to address the aforementioned challenges is still lacking. Specifically, current integration detection technologies still have significant shortcomings in the following aspects: 1) Current integration detection lacks an integrated targeted panel design for fragmented integration, resulting in incomplete detection of integration sites: Existing capture probe sets are mostly designed around a few host genes or parts of the virus, without systematically considering the fragmented integration of viruses such as AAV and HBV across the entire genome in the form of multiple fragments, missing ITR / primer binding regions, and complex rearrangements. When the virus integrates in the above forms, both traditional PCR amplification methods and existing targeted capture protocols are prone to problems such as primers failing to bind, amplified fragments being too long, or restriction enzyme sites being missing, making it difficult to detect some integrated clones, thus producing systematic false negatives and underestimation of integration sites and integration efficiency at the sample level; 2) Integration efficiency is easily underestimated: Existing methods usually use the number of detected integration sites or the number of chimeric reads as indirect indicators. However, due to the different detection sensitivities of different methods for fragmented integration, the "integration ratio" given for the same sample varies greatly under different platforms or different primers / enzyme digestion conditions. There is still a lack of a standardized technical route based on a unified panel design that can directly quantify integration efficiency from sequencing data. Summary of the Invention

[0006] To address the technical problem that existing PCR, LAM-PCR, or whole-genome sequencing (WGS) methods encounter with viruses such as HBV exhibiting "fragmented integration," incomplete primer amplification and difficulty in covering some integration fragments lead to incomplete detection of integration sites and a systematic underestimation of integration efficiency, thus affecting the safety assessment of gene therapy products. This invention addresses this issue by designing a targeted capture sequencing panel specifically for detecting viral integration sites and quantifying viral integration efficiency. This panel deploys high-density probes across the full-length virus and its typical integration hotspots to encompass multiple potential integration fragments and multi-regional integration events. Therefore, the first objective of this invention is to provide a targeted capture sequencing panel. The second objective of this invention is to provide a method for detecting viral integration sites and quantitatively evaluating integration efficiency based on targeted enrichment sequencing (TES). This method includes using the targeted enrichment sequencing panel to target and sequence samples; performing virus-host dual-reference alignment and breakpoint clustering on the enriched sequencing data to achieve comprehensive detection of integration sites; and calculating the integration efficiency by the abundance of virus-host fusion reads at each integration site relative to the total amount of virus-specific reads (or reference segment coverage). This allows for the simultaneous quantitative evaluation of integration site profiles and integration efficiency in the same sequencing experiment.

[0007] To achieve the above objectives, the present invention adopts the following technical solution: As a first aspect of the present invention, a method for detecting viral integration sites based on targeted capture sequencing includes the following steps: S1. Design a targeted capture sequencing panel that covers the full-length viral sequence and viral integration hotspots on the host genome. S2. Use the aforementioned targeted capture sequencing panel to target and capture sample DNA, construct a library, and sequence it to obtain sequencing data. S3. Align the sequencing data to the joint reference sequence of the host and virus, and extract candidate integration evidence reads; S4. Based on the alignment information of the candidate integration evidence reads on the host and viral reference sequences, the integration breakpoints are precisely located to obtain candidate breakpoint records. S5. Cluster and filter the candidate breakpoint records to obtain high-confidence integration sites.

[0008] According to the present invention, the design of the targeted capture sequencing panel described in S1 includes: S11. Analyze the homology regions between the viral vector sequence and the host genome; S12. Shield the homology region and retain the virus-specific sequence as the target for targeted capture, and design the targeted capture sequencing panel.

[0009] According to the present invention, the candidate integrated evidence reads described in S3 satisfy any of the following conditions: A. Different segments of the same read were aligned to the viral reference sequence and the host reference sequence respectively, resulting in segmented alignment across the host-virus boundary; B. In paired sequencing, one end of the read is aligned to the viral reference sequence, and the other end of the read is aligned to the host reference sequence, and the alignment quality of both ends meets the preset threshold. C. If there are obvious soft-cut or unaligned fragments on the virus or host reference sequence, a unique or high-quality alignment result can be obtained on another reference sequence after a second alignment.

[0010] Furthermore, the fine-tuning of the integration breakpoint described in S4 includes: A. For reads that satisfy condition A in S3, determine the boundary position of each aligned fragment on the read based on its CIGAR information, and map the boundary to the host genome and the viral reference sequence, respectively recording the chromosome, coordinates and chain direction of the host end breakpoint, and the reference sequence number, coordinates and chain direction of the viral end breakpoint; B. For reads that satisfy condition C in S3, align the fragment separately to the host or viral reference sequence again. If a unique or high-quality alignment is obtained, use the alignment boundary as the breakpoint position.

[0011] According to the present invention, the clustering and filtering of candidate breakpoint records described in S5 includes: S51. On the host side, using "chromosome + chain direction" as the grouping key, breakpoints whose physical location distance is within a preset window are regarded as the same host-side breakpoint cluster; S52. On the virus side, using "virus reference sequence ID + chain direction" as the grouping key, breakpoints whose virus end coordinates are located within a preset window are further merged into the same virus side breakpoint cluster. S53. Combine breakpoints that simultaneously satisfy the proximity conditions of the host side and the virus side to construct a host-virus breakpoint cluster, and regard the host-virus breakpoint cluster as a candidate integration site. S54. For each candidate integration site, count the number of independent integration evidence reads supporting the host-virus breakpoint cluster, and compare it with a preset support threshold. Only breakpoint clusters with a number of supporting reads not less than the threshold are retained as high-confidence candidate integration sites.

[0012] As a second aspect of the present invention, a method for quantitative detection of viral integration efficiency based on targeted capture sequencing includes obtaining one or more high-confidence integration sites by applying the viral integration site detection method based on targeted capture sequencing as described in any of the preceding claims; and, S6. Calculate the integration efficiency based on the integration evidence read coverage of the high-confidence integration site neighborhood and the total coverage of the target virus reads.

[0013] Furthermore, the computational integration efficiency described in S6 includes the following steps: S61. For each of the aforementioned high-confidence integration sites, calculate the integration coverage and total coverage of each high-confidence integration site; S62. Calculate the integrated coverage and total coverage for each candidate window. S63. Determine the main peak window: Select the window with the highest coverage of integrated evidence reads as the main peak window; S64. Calculate the ratio of the integrated evidence read coverage within the main peak window to the total coverage of the target virus reads, as the sample integration efficiency.

[0014] Furthermore, the computational integration efficiency described in S6 includes the following steps: The calculation of integration coverage and total coverage for each high-confidence integration site in S61 includes: For each high-confidence integration site i, the integration evidence reads coverage is calculated base-by-base using the integration evidence reads. And calculate the total coverage of the target virus reads using all reads aligned to the target virus. .

[0015] The calculation of the integrated coverage and total coverage for each candidate window as described in S62 includes: Set the window half-width w (preferably 50–100 bp), and for each high-confidence breakpoint cluster k, use... Define a candidate window at the center: Window integration coverage: Total coverage of the target virus window:

[0016] The method for determining the main peak window as described in S63 includes: Select the candidate with the highest integration coverage from all valid candidate windows: The main peak quantitative region is:

[0017] The ratio of the integrated evidence read coverage within the main peak window to the total target virus read coverage, as described in S4, is used as the sample integration efficiency. .

[0018] As a third aspect of the present invention, a targeted capture sequencing panel for detecting viral integration sites is prepared using the following steps: A. Analyze the homology regions between the viral vector sequence and the host genome; B. Shield the homology region and retain the virus-specific sequence as a target for targeted capture.

[0019] As a fourth aspect of the present invention, a targeted capture sequencing panel for quantitative detection of viral integration efficiency is prepared using the following steps: A. Analyze the homology regions between the viral vector sequence and the host genome; B. Shield the homology region and retain the virus-specific sequence as a target for targeted capture.

[0020] Furthermore, the virus-specific sequence includes the viral ITR region, promoter / regulatory element, and viral coding sequence.

[0021] As a fifth aspect of the present invention, a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method for detecting viral integration sites based on targeted capture sequencing as described in any of the preceding claims.

[0022] As a sixth aspect of the present invention, a virus integrated detection system includes: A. Sequencing module, used to acquire targeted capture sequencing data; B. An analysis module configured to perform the steps of the method for detecting viral integration sites based on targeted capture sequencing; or configured to perform the steps of the method for quantitatively detecting viral integration efficiency based on targeted capture sequencing. C. Output module, used to output a list of integration sites and / or quantitative results of integration efficiency.

[0023] The present invention provides a method for detecting integration sites and quantifying viral integration efficiency based on targeted capture sequencing. The advantages of this method are as follows: Compared to LAM-PCR methods that rely on single primers or ligation amplification, the TES Panel of this invention significantly improves the sensitivity and comprehensiveness of detecting fragmented integration through multi-site, multi-fragment targeted capture, thus more accurately reflecting the viral integration load in the host genome. This method can comprehensively and accurately assess viral integration status and efficiency, providing crucial technical support for quality control of gene therapy products, risk warning of viral infection-related diseases, and safety assessment. Attached Figure Description

[0024] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0025] Figure 1 The flowchart shows the process for determining the integration site and integration efficiency of HBV-TES.

[0026] Figure 2 The length distribution of candidate integrated evidence reads is effectively matched.

[0027] Figure 3 This shows the distribution of integration sites on different chromosomes of the host genome.

[0028] Figure 4 This represents the coverage distribution of integration sites on the viral genome.

[0029] Figure 5 The coverage curve of the candidate window (100bp) for the neighborhood of the HBV high-confidence breakpoint.

[0030] Figure 6 Comparison of integration efficiency between TES and LAM-PCR. Detailed Implementation

[0031] The present invention will be described in detail below with reference to specific embodiments and examples, thereby making the advantages and various effects of the present invention more clearly apparent. Those skilled in the art should understand that these specific embodiments and examples are for illustrative purposes only and are not intended to limit the present invention.

[0032] Throughout this specification, unless otherwise specified, the terminology used herein should be understood as having the meaning commonly used in the art. Therefore, unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. In the event of any conflict, this specification shall prevail.

[0033] Unless otherwise specified, all raw materials, reagents, instruments and equipment used in this invention can be purchased from the market or prepared by existing methods.

[0034] The specific meanings of the terms used in this invention are as follows: HBV: Hepatitis B virus.

[0035] Targeted Enrichment Sequencing (TES): A technical approach that uses probes to capture / enrich target sequences before performing high-throughput sequencing.

[0036] TES Panel / Targeted Panel: A set of probes designed for TES. This invention emphasizes coverage of the full length of the virus plus typical integration hotspots / high-risk areas to improve fragmented integration detection.

[0037] WGS (Whole-Genome Sequencing): Whole-genome sequencing has a wide coverage, but it often requires high depth for low-abundance integrated fragments, resulting in higher costs and analytical complexity.

[0038] PCR: Amplification of target fragments based on specific primers.

[0039] LAM-PCR: A linear amplification-mediated method for detecting integration sites, relying on primers and restriction enzyme digestion / ligation steps.

[0040] nrLAM-PCR: An improved version of LAM-PCR.

[0041] FASTQ: FASTQ is a text file format used to store biological sequence data and their corresponding base quality scores.

[0042] Reads: Short sequence fragments obtained in sequencing technology, used for subsequent data analysis.

[0043] read_id: A unique identifier for a sequencing read.

[0044] Barcode: Usually refers to a sample-specific short DNA sequence tag used to distinguish samples or molecules from different sources in the same sequencing reaction.

[0045] Primary alignment: The alignment record with the highest aligner score.

[0046] Secondary alignment: Substitute hits (non-fragmentation) of the same segment, flag 0x100.

[0047] MAPQ (Mapping Quality): Matching quality, reflecting the uniqueness / confidence of the match (the higher the value, the more reliable).

[0048] Viral integration: The process by which a viral genome or fragments are inserted into the host chromosome.

[0049] Integration site (IS): The specific genomic location in the host genome where viral insertion occurs (usually represented by chromosome number + coordinates + chain direction).

[0050] Integration efficiency: A quantitative indicator of "integrated viral components" in the sample relative to "total viral load / reference coverage", used to reflect the true integration load and security risk assessment.

[0051] Fragmented integration: The phenomenon that viral genomes are broken, deleted, inverted, or rearranged and then dispersed into different regions of the host as multiple fragments is an important reason why traditional primer amplification is prone to missed detection.

[0052] Minimum match length: To reduce false positives, the virus / host side must reach a certain effective match length.

[0053] RepeatMasker filtering: Uses repetitive sequence annotation to eliminate false positive evidence resulting from highly repetitive regions.

[0054] Breakpoint: The location of the connection boundary between the host sequence and the virus sequence on the read or reference.

[0055] Candidate breakpoint records: Each piece of integrated evidence read is parsed into standardized fields (readID, host chromosome / coordinates / link direction, virus reference ID / coordinates / link direction, etc.).

[0056] Coverage (depth / coverage): The number of times a certain position is covered by reads.

[0057] CIGAR (Compact Idiosyncratic Gapped Alignment Report): In genomics, CIGAR is a key field in standard genome alignment file formats such as SAM / BAM. It precisely describes, in compact string form, all alignment operations that occur sequentially, starting from the first base, when a sequenced sequence is aligned with a reference genome sequence.

[0058] Example 1 1. Design of TES Panel First, the viral vector sequence is analyzed to reduce interference from host genome homologous sequences on the targeted capture strategy. The specific method is as follows: (1) Homology analysis Sequence alignment analysis was performed between the target viral vector sequence (including but not limited to AAV, HBV, etc.) and the host reference genome to assess the level of homology between the viral sequence and the host genome. ① If the alignment results show that the viral vector sequence has no significant homologous region in the host genome, the vector sequence is considered to have high specificity, and a targeted capture probe can be designed based on its full-length sequence for subsequent enrichment and detection of integration sites; ② If the comparison results show that there are segments in the viral vector sequence that are highly homologous to the host genome (for example, the target gene expression cassette sequence is consistent with the host endogenous gene sequence), these regions will lead to a large enrichment of host genomic DNA fragments, thereby masking the real virus-host chimeric sequence signal, reducing detection sensitivity and specificity, and constituting a technical defect that interferes with detection. (2) Specific dominant region screening and probe design To overcome the technical defects caused by homology interference, this invention proposes a probe design strategy based on homology sequence shielding: before probe design, the segments in the viral vector sequence that are homologous to the host genome are automatically identified and shielded, and only virus-specific sequences (such as viral ITR regions, promoters / regulatory elements, viral coding sequences, etc.) are retained as target sites for panel design.

[0059] The optimized panel described above can significantly improve the capture efficiency and detection specificity of virus-host integration linkage sequences. It is suitable for the precise localization and quantitative analysis of viral integration sites under high-throughput sequencing (NGS) platforms, and is especially suitable for safety assessment of viral vector integration such as AAV in gene therapy products.

[0060] 2. Breakpoint identification and integration site determination based on integrated evidence The TES panel, designed and synthesized according to the above steps, is used to target hybridize and capture genomic DNA from samples, construct sequencing libraries, and sequence the DNA using a high-throughput sequencing platform (such as Illumina). The resulting raw FASTQ data is known as "raw data obtained from targeted capture sequencing (TES)". These raw data are then analyzed according to the following procedure.

[0061] (1) Sequencing data preprocessing and quality control First, the raw data obtained from targeted capture sequencing (TES) is preprocessed and quality controlled to ensure the accuracy and reliability of subsequent alignment and breakpoint resolution.

[0062] (2) Dual reference sequence alignment and extraction of candidate integrated evidence reads First, the quality-controlled filtered TES sequencing data were aligned to a joint reference sequence containing both the host genome reference sequence and the viral reference sequence. From the alignment results, reads meeting any of the following criteria were extracted as candidate "integrated evidence reads" according to pre-defined cross-host-viral boundary determination rules: ① Different segments of the same read were aligned to the viral reference sequence and the host reference sequence respectively, resulting in segmented alignments across the host-virus boundary; ② In paired-end sequencing, one end of the read aligns to the viral reference sequence, and the other end of the read aligns to the host reference sequence, with the alignment quality at both ends meeting a preset threshold; ③ There are obvious soft-clip or unaligned fragments on the virus or host reference sequence. These fragments can be uniquely or with high quality aligned to another reference sequence after a second alignment.

[0063] For the aforementioned candidate integrated evidence reads, this invention eliminates false positives caused by low-complexity sequences and repetitive sequences by setting quality control conditions such as minimum matching length, maximum mismatch ratio, and filtering of repetitive sequences, thereby obtaining a high-confidence set of candidate integrated evidence, which serves as input for subsequent breakpoint identification and integration site determination.

[0064] (3) Fine-grained localization of cross-host virus breakpoint coordinates Based on the candidate integrated evidence reads obtained in step (2), this invention performs fine-grained localization of breakpoints at cross-host-virus links using their alignment information on host and viral reference sequences. Specifically, this includes: ① For reads that are mapped to host and viral reference sequences in the form of segmented alignment, the boundary position of each aligned fragment on the read is determined according to its CIGAR information, and the boundary is mapped to the host genome and viral reference sequence. The chromosome, coordinates and strand direction of the host end breakpoint, and the reference sequence number, coordinates and strand direction of the viral end breakpoint are recorded respectively. ② For reads containing long soft-clipping or unaligned fragments, align the fragment separately to the host or viral reference sequence. If a unique or high-quality alignment is obtained, use the alignment boundary as the breakpoint location.

[0065] Through the above processing, each integrated evidence read is converted into one or more candidate breakpoint records with clear host-side coordinates and virus-side coordinates. Each record includes at least the following information: host chromosome number, host breakpoint location, host chain direction, virus reference sequence ID, virus breakpoint location, virus chain direction, and read identifier supporting the breakpoint, providing standardized input for subsequent breakpoint clustering and integration site determination.

[0066] (4) Breakpoint clustering and preliminary integration site merging For the candidate breakpoint records obtained in step (3), due to factors such as alignment algorithm errors, sequencing errors, and local minor insertions / deletions, even reads originating from the same real integration event may have "jittering" breakpoint coordinates within a range of several bases. Therefore, this invention clusters and merges candidate breakpoint records to recover the real integration sites, specifically including: ① On the host side, using "chromosome + chain direction" as the grouping key, breakpoints whose physical location distance is within a preset window are considered to be the same host-side breakpoint cluster; ② On the virus side, using "virus reference sequence ID + chain direction" as the grouping key, breakpoints whose virus-end coordinates are located within a preset window are further merged into the same virus-side breakpoint cluster; ③ For a combination of breakpoints that simultaneously satisfies the proximity conditions on both the host and virus sides, construct the corresponding “host-virus breakpoint cluster” and regard the breakpoint cluster as a candidate integration site; ④ For each candidate integration site, count the number of independent integration evidence reads supporting the breakpoint cluster and compare it with a preset support threshold. Only breakpoint clusters with a number of supporting reads not less than the threshold are retained as high-confidence candidate integration sites.

[0067] Through the above-mentioned breakpoint clustering and merging process, this invention can merge multiple "jittering" breakpoints generated by multiple reads from the same real integration event into a single integration site, thereby reducing site fragmentation and redundant statistics caused by coordinate jitter, and providing a stable site-level input for accurate quantification of subsequent integration efficiency.

[0068] The breakpoint clusters obtained after the above quality control are used as the "high-confidence integration site set" determined by this invention, providing a reliable basis for subsequent quantitative integration efficiency and functional annotation.

[0069] 3. Quantitative assessment of integration efficiency Given that target viruses can exhibit fragmented integration characteristics within the host genome, integration evidence reads often show localized spike-like enrichment on the viral genome side. In this context, cumulative quantification of multiple low-support or scattered breakpoints easily introduces noise and capture fluctuations, reducing the stability of cross-sample comparisons. Therefore, this invention further proposes a quantitative assessment strategy for integration efficiency based on the "main integration peak (highest integration peak)," that is, selecting only the neighborhood of the breakpoint with the highest peak in the integration evidence coverage curve as the quantification region, with the integration peak intensity in this region representing the main integration contribution of the sample. This strategy is particularly suitable for samples where fragmented integration is widespread and there are many breakpoints with significant differences in strength.

[0070] The specific steps are as follows: (1) Calculate the integration coverage and total coverage for each high-confidence integration site. For each high-confidence integration site i, the integration evidence read coverage is calculated base-by-base using only the integration evidence reads. And calculate the total coverage of the target virus reads using all reads aligned to the target virus. .

[0071] (2) Calculate the integrated coverage and total coverage for each candidate window. Set the window half-width w (preferably 50–100 bp), and for each high-confidence breakpoint cluster k, use... Define a candidate window at the center: Window integration coverage: Total coverage of the target virus window:

[0072] (3) Determine the main peak window Select the candidate with the highest integration coverage from all valid candidate windows: The main peak quantitative region is:

[0073] (4) Definition of integration efficiency of local proportion of main peak The local proportion within the main peak window is used as the sample integration efficiency: Where X represents each genomic position within the main peak window, the molecular statistic represents the coverage of integrated evidence reads, and the denominator represents the coverage of all viral reads.

[0074] This scheme limits the quantitative anchor point to the neighborhood of the high-confidence breakpoint cluster and selects only the main peak region with the strongest integration evidence for local proportional normalization. This can highlight the main integration contribution of the sample in the context of fragmented integration and reduce the interference of multiple weak peaks, background fluctuations and uneven local capture on the integration efficiency assessment. At the same time, since the main peak comes from the global capture of TES and the clustering results of high-confidence breakpoints, it avoids the limitation of LAM-PCR, which relies on a single amplification site and is difficult to determine the main peak at the viral genome scale.

[0075] The following will provide a detailed description of the detection method for viral integration sites based on targeted capture sequencing and the quantitative detection method for integration efficiency of this application, in conjunction with embodiments and experimental data.

[0076] Example 2 The main objective of Example 2 is to verify the existence of fragmented integration of HBV into the host genome. This means that the viral genome is not inserted as a single, complete fragment, but rather as multiple fragments of varying lengths and origins integrated into multiple host sites. Based on this phenomenon, it is further explained that when fragmented integration occurs, traditional LAM-PCR, which focuses on a single site, is more susceptible to primer position and amplification bias, only capturing local or single breakpoint evidence. It struggles to cover all fragmented integration events in the viral genome and is therefore unsuitable for objectively and comparablely quantitatively assessing the overall "integration efficiency." In contrast, the global capture strategy employed in this invention can systematically capture various types of fragmented integration evidence at the viral genome scale, making it more suitable for the comprehensive identification of fragmented integration events and the quantitative analysis of integration efficiency.

[0077] To further verify the feasibility and accuracy of the proposed method for detecting viral integration sites and quantitatively evaluating integration efficiency based on targeted capture sequencing (TES), Example 2 selected one hepatitis B-related hepatocellular carcinoma (HCC) tissue sample. The HBV-TES Panel designed in Example 1 was used for library construction and high-throughput sequencing. Through steps such as host-virus dual reference alignment, integration evidence read extraction, fine identification of cross-host-virus breakpoints, and breakpoint clustering, high-confidence HBV integration sites in the sample were systematically identified.

[0078] 1. Data acquisition and preprocessing Genomic DNA was extracted from a surgical specimen of HBV-related HCC. A capture library was constructed using the HBV-TES Panel designed in Example 1, and paired-end sequencing was performed on an Illumina platform (PE150). The nucleotide sequences of HBV-TESPanel are detailed in SEQ ID No. 1 to SEQ ID No. 144 in the sequence listing.

[0079] Raw data were obtained through sequencing. Quality control software was used to identify and remove adapter sequences from the raw reads of each sample, and low-quality bases at both ends of the reads were cleaved. The results are shown in Table 1.

[0080] Table 1. Statistical results of TES sequencing data quality control

[0081] 2. Dual reference sequence alignment and extraction of candidate integrated evidence reads In this embodiment, BWA was used to align the quality-controlled filtered TES sequencing data to a combined reference sequence assembled from the human genome hg38 and HBV reference sequences, and to remove duplicates, retaining primary, supplementary, and secondary alignment information. From the alignment results, reads meeting any of the following conditions were extracted as candidate "integrated evidence reads": ① Split-reads: Different segments of the same read are aligned to the host reference sequence and the HBV reference sequence respectively. There are obvious split alignments or soft splicing structures in CIGAR. ② Discordant pairs: In paired-end sequencing, one end of the read is aligned to the host genome with high quality, and the other end of the read is aligned to the HBV reference sequence with high quality. Both ends of the read have an alignment quality score ≥ Q20 and an effective match length ≥ 20 bp. ③ Soft-clipping re-alignment reads: There is a soft-clipping fragment of ≥25 bp in length on the host or HBV reference sequence. After a second alignment of this fragment, a unique or high-quality alignment can be obtained on another reference sequence.

[0082] To ensure the reliability of candidate integration evidence, this embodiment further sets a minimum matching length of ≥25 bp and an alignment quality score (MAPQ) of ≥20. In addition, RepeatMasker annotation is used to remove alignments that fall into highly repetitive regions. This set of high-confidence candidate integration evidence serves as the input for subsequent fine-grained localization of cross-host-virus breakpoints and clustering of integration sites.

[0083] In this embodiment, the type and quantity distribution of the above-mentioned candidate integrated evidence reads are shown in Table 2.

[0084] The results in Table 2 show that segmented and paired cross-source reads account for the majority, while soft-cut re-alignment reads serve as supplementary evidence to further improve the sensitivity of breakpoint identification.

[0085] The matching lengths of candidate integrated evidence reads are shown in Figure 2. Figure 2 The results showed that the effective match length of most reads was much higher than 25 bp.

[0086] Table 2. Classification and Quantity Statistics of Candidate Integrated Evidence Reads in this Embodiment

[0087] 3. Cross-host – fine-grained localization of virus breakpoint coordinates In this embodiment, each candidate integration evidence read obtained in step (2) is analyzed, and the breakpoints at the cross-host-virus connection are precisely located based on the alignment information of the reads on the host genome and the viral reference sequence.

[0088] First, for reads with split-read characteristics, their alignment start and end positions and corresponding CIGAR information on the host and viral reference sequences are read to determine the boundaries of each aligned fragment on the read, and these boundaries are mapped to the host and viral reference coordinate systems: On the host side, record the chromosome where the breakpoint is located, its specific coordinates, and the chain direction information; On the viral side, the viral reference sequence ID, specific nucleotide position, and chain direction information at the breakpoint are recorded.

[0089] Secondly, for reads containing long soft-clipping or unaligned fragments, these fragments are extracted separately and aligned again using the same joint reference sequence described above. If the fragment achieves a unique or high-quality alignment with the host or viral reference sequence, the boundary between the fragment and the original aligned fragment is taken as the cross-host-viral breakpoint location, and the corresponding host-end and viral-end coordinates and chain direction information are recorded.

[0090] Through the above processing, each integrated evidence read is converted into a standardized "candidate breakpoint record". Each candidate breakpoint record includes at least the following fields: host chromosome number, host breakpoint coordinates, host chain direction, viral reference sequence ID, viral breakpoint coordinates, viral chain direction, and read identifier supporting that breakpoint. See Table 3 for an example.

[0091] Table 3 Examples of Candidate Breakpoint Records

[0092] 4. Breakpoint clustering and preliminary integration site merging In this embodiment, the candidate host-virus breakpoint records obtained in step (3) are clustered based on proximity on both the host and virus sides, and candidate integration sites are constructed based on this.

[0093] First, on the host side, candidate breakpoint records are grouped according to "chromosome + chain direction," and a preset clustering window (10bp) is set. Breakpoints whose positions fall within the same window interval are grouped into the same host-side breakpoint cluster. For each host-side breakpoint cluster, its representative breakpoint coordinates are recorded as the host-side candidate integration positions for that cluster.

[0094] Secondly, on the virus side, candidate breakpoint records are grouped according to "HBV reference sequence ID + chain direction". A similar preset window is used to merge breakpoints with adjacent virus-side coordinates into the same virus-side breakpoint cluster, and the representative breakpoint coordinates of each virus-side breakpoint cluster are determined.

[0095] Based on this, host-side breakpoint clusters and virus-side breakpoint clusters are matched accordingly: for candidate breakpoint records that simultaneously satisfy the condition that "the host end falls within a host-side breakpoint cluster" and "the virus end falls within a virus-side breakpoint cluster," they are assigned to the same "host-virus breakpoint cluster," and this cluster is considered a candidate integration site. For each candidate integration site, the number of independent integration evidence reads it contains is counted, and this number is used as the support index for that integration site.

[0096] In this embodiment, based on a pre-set support threshold, only host-virus breakpoint clusters with a number of supporting reads not less than the threshold are retained as high-confidence candidate integration sites; breakpoint clusters with support below the threshold are marked as low-confidence sites and are not included in subsequent quantitative analysis of integration efficiency.

[0097] After the above clustering and screening, a total of 445 candidate integration sites were identified in this embodiment (see Table 4). After filtering based on the number of supporting reads, alignment quality, and breakpoint consistency, 214 high-confidence integration sites were finally obtained. These high-confidence integration sites are distributed on 24 different chromosomes, and the number of integration sites varies on different chromosomes, with some chromosomes detecting a relatively large number of integration sites (see Figure 3).

[0098] Table 4. Statistical results of support for candidate integration sites in the HCC-1 sample.

[0099] From the distribution of viral integration breakpoints, HBV integration breakpoints exhibit a multi-peak distribution along the viral genome, accounting for 65% of the total length and covering multiple functional segments, not limited to a single region. Breakpoints are relatively enriched in the preS, S, and X gene regions, but stable integration signals can also be detected in the core region and polymerase coding region, indicating that this method can effectively capture scattered and fragmented viral integration events (see Figure 4).

[0100] In summary, the method described in this application demonstrates that HBV may participate in host-virus rearrangement in a multi-breakpoint, cross-regional manner, supporting the conclusion that fragmented integration phenomena objectively exist in clinical samples. Furthermore, it shows that the global capture and breakpoint clustering strategy based on TES can stably integrate candidate breakpoint records and effectively recover real integration events at the viral genome scale, thus providing a reliable foundation for subsequent quantitative assessment of integration efficiency in the context of fragmented integration.

[0101] Example 3 The main objective of Example 3 is to further verify the feasibility, stability, and comparability of the proposed TES-based viral integration efficiency quantitative assessment strategy, building upon the high-confidence viral integration sites obtained in Example 2. Specifically, this example uses TES data from the same sample as a basis and limits the quantitative anchor point to the neighborhood of the high-confidence viral breakpoint cluster within the viral genome coordinate system. Furthermore, it selects only the main peak region with the strongest integration evidence for local proportional normalization to establish a sample-level integration efficiency index. This strategy can highlight the main integration contribution of the sample in the context of fragmented integration, reducing the interference of multiple weak peaks, background fluctuations, and uneven local capture on integration efficiency assessment. Simultaneously, since the determination of the main peak originates from the global capture and high-confidence breakpoint clustering results of TES, this invention avoids the limitations of traditional LAM-PCR, which relies on a single amplification site and struggles to objectively determine the main peak at the viral genome scale. This embodiment compares the integration efficiency results obtained by the present invention with the integration detection results of the traditional LAM-PCR method to demonstrate that the method of the present invention can more objectively reflect the viral integration load in the context of fragmented integration, and provide a reliable quantitative basis for the safety evaluation of gene therapy products and the risk assessment of HBV-related tumors.

[0102] 1) Calculate the integration coverage and total coverage for each high-confidence integration site. For each high-confidence integration site i in Example 1, the integration evidence reads coverage is calculated base-by-base using only the integration evidence reads. And calculate the total HBV read coverage using all reads aligned to HBV. .

[0103] 2) Calculate the integrated coverage and total coverage for each candidate window. Set the window half-width w (100 bp), and for each high-confidence breakpoint cluster k, use... Define a candidate window at the center: Window integration coverage: Total HBV coverage: And draw a coverage curve ( Figure 5 ).

[0104] 3) Determine the main peak window Select the candidate with the highest integration coverage from all valid candidate windows: The main peak quantitative region is: .

[0105] 4) Definition and integration efficiency of the main peak's local proportions The local proportion within the main peak window is used as the sample integration efficiency: .

[0106] In this embodiment, the high-confidence viral side breakpoints cover multiple regions of the HBV genome and exhibit local clustering (see...). Figure 4 By defining candidate windows only within the neighborhood of high-confidence breakpoint clusters and selecting the main peak window with the highest coverage of integrated evidence, the quantitative anchor point for integration efficiency can be locked onto the main integration hotspot region with the highest support and most stable peak shape. Furthermore, the comparison results of integration efficiency indices show that the integration efficiency index obtained based on the local proportion of the TES main peak is generally higher than the corresponding index obtained based on a single LAM-PCR site (see...). Figure 6 This indicates that, in the context of fragmented integration breakpoints and the coexistence of strong and weak peaks, single amplification site methods are easily limited by site selection and amplification bias, making it difficult to reflect the main integration contribution at the viral genome scale. In summary, the global capture of TES combined with the main peak local normalization strategy can better highlight the main integration signal in the context of fragmented integration and reduce the impact of weak peaks and background fluctuations on overall quantification, thus verifying the feasibility and stability of the quantitative assessment method for integration efficiency of this invention.

[0107] In summary, this invention provides a systematic solution to two major pain points in existing integration detection technologies: First, it addresses the problems of existing targeted capture panel designs lacking fragmented integration targeting, failing to fully cover multi-fragment integration across the entire genome of viruses such as AAV and HBV, missing ITR / primer binding regions, and complex rearrangements. This avoids missed detection of integrated clones caused by primer binding failure, abnormal amplified fragments, or missing restriction enzyme sites in traditional PCR and existing targeted capture schemes, fundamentally reducing the risk of systemic false negatives in sample detection and improving the comprehensiveness of integration site detection. Second, it overcomes the problem that existing methods rely on indirect indicators (such as the number of integration sites and the number of chimeric reads), leading to inaccurate quantitative integration efficiency and poor comparability of results under different platforms / experimental conditions. Through a unified panel design, it constructs a standardized technical route that can directly quantify integration efficiency from sequencing data, achieving a dual breakthrough in the comprehensiveness and quantitative accuracy of integration detection.

[0108] Finally, it should be noted that the terms “comprising,” “including,” or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0109] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0110] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A method for detecting viral integration sites based on targeted capture sequencing, characterized in that, Includes the following steps: S1. Design a targeted capture sequencing panel that covers the full-length viral sequence and viral integration hotspots on the host genome; S2. Use the aforementioned targeted capture sequencing panel to target and capture sample DNA, construct a library, and sequence it to obtain sequencing data. S3. Align the sequencing data to the joint reference sequence of the host and virus, and extract candidate integration evidence reads; S4. Based on the alignment information of the candidate integration evidence reads on the host and viral reference sequences, the integration breakpoints are precisely located to obtain candidate breakpoint records. S5. Cluster and filter the candidate breakpoint records to obtain high-confidence integration sites.

2. The method for detecting viral integration sites based on targeted capture sequencing as described in claim 1, characterized in that, The design of the targeted capture sequencing panel described in S1 includes: S11. Analyze the homology regions between the viral vector sequence and the host genome; S12. Shield the homology region and retain the virus-specific sequence as the target for targeted capture, and design the targeted capture sequencing panel.

3. The method for detecting viral integration sites based on targeted capture sequencing as described in claim 1, characterized in that, The candidate integrated evidence reads described in S3 satisfy any of the following conditions: A. Different segments of the same read were aligned to the viral reference sequence and the host reference sequence respectively, resulting in segmented alignment across the host-virus boundary; B. In paired sequencing, one end of the read is aligned to the viral reference sequence, and the other end of the read is aligned to the host reference sequence, and the alignment quality of both ends meets the preset threshold. C. If there are obvious soft-cut or unaligned fragments on the virus or host reference sequence, a unique or high-quality alignment result can be obtained on another reference sequence after a second alignment.

4. The method for detecting viral integration sites based on targeted capture sequencing as described in claim 3, characterized in that, The fine-tuning of the integration breakpoint described in S4 includes: A. For reads that meet condition A, determine the boundary position of each aligned fragment on the read based on its CIGAR information, and map the boundary to the host genome and the viral reference sequence, respectively recording the chromosome, coordinates and chain direction of the host end breakpoint, and the reference sequence number, coordinates and chain direction of the viral end breakpoint; B. For reads that meet condition C, align the fragment separately to the host or viral reference sequence. If a unique or high-quality alignment is obtained, use the alignment boundary as the breakpoint position.

5. The method for detecting viral integration sites based on targeted capture sequencing as described in claim 1, characterized in that, The clustering and filtering of candidate breakpoint records described in S5 includes: S51. On the host side, using "chromosome + chain direction" as the grouping key, breakpoints whose physical location distance is within a preset window are regarded as the same host-side breakpoint cluster; S52. On the virus side, using "virus reference sequence ID + chain direction" as the grouping key, breakpoints whose virus end coordinates are located within a preset window are further merged into the same virus side breakpoint cluster. S53. Combine breakpoints that simultaneously satisfy the proximity conditions of the host side and the virus side to construct a host-virus breakpoint cluster, and regard the host-virus breakpoint cluster as a candidate integration site. S54. For each candidate integration site, count the number of independent integration evidence reads supporting the host-virus breakpoint cluster, and compare it with a preset support threshold. Only breakpoint clusters with a number of supporting reads not less than the threshold are retained as high-confidence candidate integration sites.

6. A quantitative detection method for viral integration efficiency based on targeted capture sequencing, characterized in that, This includes obtaining one or more high-confidence integration sites using the viral integration site detection method based on targeted capture sequencing as described in any one of claims 1-5; as well as, S6. Calculate the integration efficiency based on the integration evidence read coverage of the high-confidence integration site neighborhood and the total coverage of the target virus reads.

7. The method for quantitative detection of viral integration sites and integration efficiency based on targeted capture sequencing as described in claim 1, characterized in that, The computational integration efficiency described in S6 includes the following steps: S61. For each of the aforementioned high-confidence integration sites, calculate the integration coverage and total coverage of each high-confidence integration site; S62. Calculate the integrated coverage and total coverage for each candidate window; S63. Determine the main peak window: Select the window with the highest coverage of integrated evidence reads as the main peak window; S64. Calculate the ratio of the integrated evidence read coverage within the main peak window to the total coverage of the target virus reads, as the sample integration efficiency.

8. A targeted capture sequencing panel, characterized in that, Including but not limited to those used for the detection of viral integration sites and the quantitative detection of viral integration efficiency, they are prepared using the following steps: A. Analyze the homology regions between the viral vector sequence and the host genome; B. Shield the homology region and retain the virus-specific sequence as a target for targeted capture.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the detection method as described in any one of claims 1-5, or implements the quantitative detection method as described in any one of claims 6-7.

10. A virus integrated detection system, characterized in that, include: A. Sequencing module, used to acquire targeted capture sequencing data; B. An analysis module, configured to perform the steps of the detection method as described in any one of claims 1-5; or configured to perform the steps of the quantitative detection method as described in any one of claims 6-7; C. Output module, used to output a list of integration sites and / or quantitative results of integration efficiency.

Images