Molecular tag-based analysis method and device for amplicon sequencing data, electronic equipment and storage medium

Abstract

The present application relates to the technical field of bioinformatics, and particularly relates to a molecular tag-based amplification event level sequencing data analysis method and device, electronic equipment and storage medium. The method analyzes sequencing data of amplification products carrying resolvable molecular tags: screening complete tag reads, constructing a molecular tag family based on the complete tag reads and determining a family consensus based on double-end independent fragments; screening a double-end paired tag family and a double-end consistent candidate state family from the family with a clear consensus state; determining a state analysis result of a target region or a target site based on the relationship between the number of the double-end consistent candidate state family and the total number of the double-end paired tag family. The present application can reduce the influence of non-complete tag reads, non-consensus families and double-end state conflicts on the analysis result.

Description

Technical Field

[0001] This invention relates to the field of bioinformatics technology, specifically to a method, apparatus, electronic device, and storage medium for analyzing molecular tag-based amplification event-level sequencing data. Background Technology

[0002] Amplified sequencing data are typically obtained through high-throughput sequencing of amplified products related to the target region. This type of data is characterized by concentrated target regions, a large number of reads, and ease of in-depth analysis of target regions or sites, and is therefore widely used for the analysis of target regions or sites.

[0003] In existing technologies, a common analysis method is the read layer ratio method, which analyzes the ratio between the number of candidate reads and the number of reads covered by the target region. However, amplified sequencing data may contain duplicate reads generated by amplification, as well as incomplete tag reads with missing tags, incomplete tags, insufficient tag quality, tags that can only be resolved at one end, or whose pairwise pairing relationship cannot be determined. If these reads are directly included in the statistics of the number of candidate reads or the number of reads covered by the target region, duplicate counting, statistical denominator contamination, or background noise amplification may occur, thus affecting the accuracy of the analysis results.

[0004] To mitigate the impact of duplicate reads, existing technologies have proposed statistical methods based on molecular tag families. These methods group reads with the same or similar molecular tags into the same molecular tag family and analyze them based on the number of candidate state tag families and the total number of tag families. However, within the same molecular tag family, different reads may support a reference state, a candidate state, or other states. When the highest supported state within a family does not meet the preset consensus conditions, the number of supported states is the same, or the supported state of reads within a family cannot be determined, the molecular tag family may belong to a non-consensus family or an undecidable family.

[0005] Furthermore, the first and second read segments of the same independent fragment may support different states, thus creating a conflict between the two ends of the state. If families with no consensus, conflicting states, and undecidable states are not separately labeled or excluded, and are directly included in the total number of candidate state families or label families, misjudgments may still occur.

[0006] Conventional paired-end sequencing analysis typically performs pairing or consistency checks on the first and second reads at the read level. However, it may not combine molecular tag integrity, molecular tag family merging, family consensus status, and handling of abnormal families to form evidence of paired-end consistency candidate status at the molecular tag family or amplification event level, thus affecting the accuracy of the analysis results. Summary of the Invention

[0007] To address the technical problem of insufficient accuracy in amplification sequencing data state analysis caused by the inclusion of incomplete tag reads, consensus-less families, undeterminable families, and paired-end state conflicts in the statistics in existing technologies, this invention provides a molecular tag-based amplification event-level sequencing data analysis method, apparatus, electronic device, and computer-readable storage medium.

[0008] In a first aspect, the present invention provides a method for analyzing amplification event-level sequencing data based on molecular tags. The amplification sequencing data is obtained by sequencing amplification products carrying resolvable molecular tags, and includes the following steps: (1) Determine each relevant read of the target region or target site in the amplified sequencing data and the molecular tag sequence in each relevant read, wherein each relevant read consists of a first end read and a second end read with a corresponding relationship; based on the molecular tag sequence, screen complete tag reads from each relevant read; (2) Construct molecular tag families based on each complete tag reading segment; within each molecular tag family, determine the first end reading segment and the second end reading segment with corresponding relationship as the same independent segment; use the independent segment as the statistical unit to perform family consensus determination on each molecular tag family, determine the consensus state of each molecular tag family, and screen out the determinable double-end paired tag families from the molecular tag families with clear consensus state. (3) Filter out the two-end consistent candidate state families from the families of the two-end consistent labels that can be determined, and determine the number of the two-end consistent candidate state families; determine the state analysis results of the target region or target site based on the number of the two-end consistent candidate state families.

[0009] In conjunction with the first aspect mentioned above, among some possible implementation methods, the methods for filtering out complete tagged segments from various relevant segments include: Based on the preset molecular tag structure, the tag-bearing end in the first and second end of each relevant read segment is determined, and the comprehensive attribute information of the molecular tag of the tag-bearing end is obtained. For any set of related reading segments with corresponding relationships, based on the comprehensive attribute information of the molecular tag, it is determined whether the tag carrier end of the molecular tag that needs to be parsed according to the preset molecular tag structure meets the preset tag integrity condition. If all the tag carrier ends meet the preset tag integrity condition, then determine whether the first end reading segment and the second end reading segment meet the preset valid pairing condition; if they do, then divide the relevant reading segments into complete tag reading segments.

[0010] In conjunction with the first aspect mentioned above, among some possible implementation methods, the methods for constructing molecular tag families include: The molecular tag sequence and fragment feature information of each complete tag read are combined to obtain the molecular identifier of each complete tag read. Complete tag reads with the same molecular identifier are grouped into the same molecular tag family.

[0011] In conjunction with the first aspect mentioned above, among some possible implementation methods, the methods for filtering out families of determinate paired tags include: For any molecular tag family with a clear consensus, determine whether each independent segment is a determinate double-end paired independent segment. If at least one independent segment is a determinate double-end paired independent segment, then the molecular tag family is determined as a determinate double-end paired tag family.

[0012] In conjunction with the first aspect mentioned above, among some possible implementations, methods for determining whether each independent segment is a determinate paired independent segment include: For any independent segment, determine whether the first end read segment and the second end read segment of the independent segment can determine the supported state; if both the first end read segment and the second end read segment can determine the supported state, and the first end read segment and the second end read segment support the same state, then the independent segment is determined to be a pairwise independent segment.

[0013] In conjunction with the first aspect mentioned above, among some possible implementations, methods for selecting candidate state families with consistent outcomes at both ends include: For any determinate dual-end paired tag family, determine whether the consensus state of the determinate dual-end paired tag family is a candidate state; When the consensus state of the determinate double-end pairing tag family is a candidate state, determine whether the first end read segment and the second end read segment of each determinate double-end pairing independent segment in the determinate double-end pairing tag family both support the same candidate state. If at least one determinates a pair of independent segments, both the first and second end read segments support the same candidate state, then the determinate pair of paired tags family is determined as a pairwise consistent candidate state family.

[0014] In conjunction with the first aspect mentioned above, among some possible implementation methods, the methods for determining the state analysis results of the target region or target site include: Determine the total number of determinate paired tag families and the number of paired consistent candidate state families; The support index for double-end consistent amplification events is calculated based on the total number of determinate double-end paired tag families and the number of double-end consistent candidate state families. Based on the dual-end consistent amplification event support index, it is determined whether the target region or target site meets the support conditions for the corresponding candidate state.

[0015] In conjunction with the first aspect above, in some possible implementations, the method for determining whether the target region or target site meets the support conditions for the corresponding candidate state includes: comparing the number of two-end consistent candidate state families with a preset number threshold, and comparing the two-end consistent amplification event support index with a preset support index threshold; when the number of two-end consistent candidate state families is greater than or equal to the preset number threshold, and the two-end consistent amplification event support index is greater than or equal to the preset support index threshold, it is determined that the target region or target site meets the support conditions for the corresponding candidate state.

[0016] In conjunction with the first aspect mentioned above, among some possible implementation methods, the methods for determining the state analysis results of the target region or target site also include: From the families of identifiable paired labels, select high-confidence label families and high-confidence candidate state families, and determine the total number of high-confidence label families and the number of high-confidence candidate state families, respectively. The high-confidence candidate state support index is calculated based on the total number of high-confidence label families and the number of high-confidence candidate state families. The state analysis results were reviewed based on the high-confidence candidate state support index.

[0017] In conjunction with the first aspect mentioned above, among some possible implementation methods, the methods for verifying the state analysis results include: Based on the two-end consistent amplification event support index and the high-confidence candidate state support index, the relative deviation between the two is calculated. Based on the degree of relative deviation, determine whether the state analysis results are correct.

[0018] Secondly, the present invention provides a molecular tag-based amplification event-level sequencing data analysis device, including an input module and a core analysis module; The core analysis module is used to obtain analysis results based on the received data using the molecular tag-based amplification event-level sequencing data analysis method of this invention.

[0019] Thirdly, the present invention provides an electronic device, including a processor and a memory, wherein the memory stores a computer program, and when the processor executes the computer program, it implements the molecular tag-based amplification event-level sequencing data analysis method of the present invention.

[0020] Fourthly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the present invention implements the molecular tag-based amplification event-level sequencing data analysis method.

[0021] This invention offers the following advantages: It filters complete tag reads based on molecular tag integrity and valid end-to-end pairing relationships, and constructs molecular tag families for event-level analysis solely based on these complete tag reads. This reduces the impact of missing tags, incomplete tags, insufficient tag quality, and reads with undeterminable end-to-end pairing relationships on the analysis results. Furthermore, this invention identifies corresponding first-end and second-end reads as independent fragments and uses these independent fragments as statistical units for family consensus determination, thereby reducing the impact of amplified duplicate reads on the statistical results. Furthermore, this invention requires that both the first-end and second-end reads of determinable end-to-end paired independent fragments can determine the supported state and support the same state, thus eliminating misjudgments caused by end-to-end state conflicts. Attached Figure Description

[0022] Figure 1 This is a flowchart of a molecular tag-based amplification event-level sequencing data analysis method according to an embodiment of the present invention; Figure 2 This is a schematic diagram of the molecular tag in the primer-related region and the amplification product in an embodiment of the present invention; Figure 3 This is a schematic diagram illustrating the hierarchical structure of the complete tag read segment set and the incomplete tag read segment set in an embodiment of the present invention; Figure 4 This is a schematic diagram illustrating the consensus determination and family status classification of molecular tag families in an embodiment of the present invention. Figure 5 This is a schematic diagram illustrating the screening of a pairwise matched tag family according to an embodiment of the present invention; Figure 6 This is a schematic diagram illustrating the screening of two-end consistent candidate state families and the calculation of the two-end consistent amplification event support index in an embodiment of the present invention. Figure 7 This is a schematic diagram of the structure of the molecular tag-based amplification event-level sequencing data analysis device according to an embodiment of the present invention; Figure 8 This is a schematic diagram of the structure of an electronic device according to an embodiment of the present invention. Detailed Implementation

[0023] To clearly illustrate the technical features of this solution, the invention will be described in detail below through specific embodiments and in conjunction with the accompanying drawings.

[0024] Method Implementation Examples: A flowchart of a molecular tag-based amplification event-level sequencing data analysis method of the present invention is shown below. Figure 1 As shown, it includes the following steps: Step 1: Data Acquisition and Segment Layering.

[0025] (1.1) Obtain amplification and sequencing data containing molecular tags and preset sequence structure reference information.

[0026] In this embodiment, the molecularly tagged amplification sequencing data is obtained from sequencing the amplification products. The amplification products carry resolvable molecular tags. The molecular tag sequence enters the amplification product through primers, primer extension regions, primer-associated adapter regions, amplification product end regions, or combinations thereof, and is parsed in subsequent sequencing reads. A schematic diagram of the structure of the primer-associated region molecular tag and the amplification product is shown below. Figure 2 As shown.

[0027] In this embodiment, the preset sequence structure reference information includes at least one of the following: target region (or target site, taking the target region as an example in this embodiment), reference sequence information, primer structure information, or molecular tag structure information.

[0028] Reference sequence information refers to the known standard genome sequence or reference gene sequence corresponding to the target detection region. The reference sequence information serves as a template for aligning sequencing reads. By aligning the basic read with the reference sequence, the alignment position of the basic read on the reference sequence is determined. According to the target region-related read screening rules described in step (1.2), relevant reads belonging to the target region are screened, while non-target region reads, non-specific amplified reads, or low-confidence alignment reads are filtered out.

[0029] (1.2) Preprocess the amplified sequencing data containing molecular tags to obtain a basic read set, and identify relevant reads of the target region or target site from the basic read set.

[0030] In this embodiment, the preprocessing includes one or more of the following: connector removal, low-quality base processing, short read filtering, paired-end read matching check, alignment processing, target region read screening, effective pair filtering, alignment quality filtering, base quality statistics, and duplicate information marking.

[0031] Among these methods, adapter removal can be performed using adapter sequence matching or local alignment; low-quality base processing can be performed using end trimming, sliding window trimming, or filtering methods based on Phred quality values; short read filtering can be performed based on the effective length of the processed reads; paired read matching checks can be performed based on read names, pairing identifiers, or read types; alignment processing can be performed using short read sequence alignment methods, comparing the base read with the reference sequence; effective pairing filtering can be performed based on paired read pairing identifiers, alignment direction, alignment position, inserted fragment range, or target region correspondence; alignment quality filtering can be performed based on whether MAPQ reaches a preset alignment quality threshold; and duplication information marking can be performed based on read names, molecular tag sequences, alignment positions, or paired pairing relationships.

[0032] After the alignment process, an aligned read file is generated, and a molecular tag sequence field, as well as optional primer pairing information, tag type information, primer information, or read classification information are written into the read file.

[0033] The method for identifying target region-related reads from the basic read set includes: filtering reads corresponding to the target region based on at least one of the following: the overlap relationship between the read alignment coordinates and the target region coordinates, the primer sequence matching relationship, the read alignment direction, the range of inserted fragments, the alignment quality, or the target region identifier.

[0034] Specifically, the overlap between the alignment position of the read and the target region can be used to determine whether the read corresponds to the target region. Reads whose alignment positions do not have a preset overlap with the target region can be identified as non-target region reads. Whether a read is a specific amplification read of the target region can be determined based on at least one of the following: target-specific primer sequence matching, read alignment position, read alignment direction, inserted fragment range, or target region correspondence. Reads that do not meet the preset target-specific primer matching conditions or target region correspondence conditions can be identified as non-specific amplification reads. The reliability of the alignment results can be determined based on the read alignment quality. When the alignment quality (MAPQ) of the read is lower than a preset alignment quality threshold, it can be identified as a low-reliability aligned read. Non-target region reads, non-specific amplification reads, and low-reliability aligned reads are not included in the target region-related read set.

[0035] The preset alignment quality threshold can be determined based on the sequencing platform, read length, sequence alignment method used, sequence complexity of the target region, negative sample analysis results, or historical validation data.

[0036] In this embodiment, the target region-related read segment set includes first-end read segments and second-end read segments. First-end read segments and second-end read segments that can establish a correspondence based on read segment names or pairing identifiers form double-end read segment pairs; read segments that fail to form valid double-end pairs or whose only one end can be effectively parsed are classified into the incomplete tag read segment set in subsequent steps.

[0037] (1.3) Analyze the molecular tag sequences in the relevant read segments of the target region and determine the comprehensive attribute information of the molecular tags.

[0038] For any target region related read segment, according to the preset molecular tag structure, the tag-bearing ends in the first and second read segments are determined, and molecular tag parsing is performed on the tag-bearing ends to obtain the molecular tag sequence and its comprehensive molecular tag attribute information.

[0039] In this embodiment, the comprehensive attribute information of the molecular tag includes at least one of the following: tag position, tag orientation, tag length, tag base quality, undetermined base ratio, relative positional relationship between the tag and the target-specific primer region or target region, tag type information, and read pairing information.

[0040] Different sequencing experiments can employ either a single-tag or dual-tag design. For a single-tag design, based on a pre-defined molecular tag structure, the molecular tag is parsed from the pre-defined tag-bearing end of the first or second read. The other read does not need to contain a molecular tag, but it is used to determine the pairwise pairing relationship and state support status. For a dual-tag design, the molecular tag is parsed from the pre-defined positions of the first and second reads, respectively, and the comprehensive attribute information of the molecular tags at both ends is obtained.

[0041] When a molecular tag sequence is not obtained at the preset tag position, or when the obtained molecular tag sequence does not meet at least one of the preset length, base quality, undetermined base ratio, or relative positional relationship, the corresponding tag missing, tag incomplete, tag quality insufficient, or tag structure abnormality identifier is recorded.

[0042] (1.4) Based on the molecular tag comprehensive attribute information of the first end reading segment and the second end reading segment, determine whether the relevant reading segments of the target region meet the preset tag integrity condition, and determine whether the first end reading segment and the second end reading segment meet the preset effective pairing condition. Based on the judgment result, divide the relevant reading segments of the target region into a complete tag reading segment set and an incomplete tag reading segment set.

[0043] A hierarchical diagram of the complete tag read set and the incomplete tag read set is shown below. Figure 3 As shown.

[0044] For any read segment related to a target region, methods for classifying it into the complete label read segment set or the incomplete label read segment set include: ①Based on the molecular tag comprehensive attribute information of the first and second end reading segments of the relevant reading segment, and according to the preset molecular tag structure, determine whether the tag carrier end of the molecular tag to be parsed meets the preset tag integrity condition; if the tag carrier ends all meet the preset tag integrity condition, then execute step (2); otherwise, divide the relevant reading segment into the incomplete tag reading segment set.

[0045] Specifically, for a single-label design, it is determined whether the preset single label carrier end meets the preset label integrity condition; for a dual-label design, it is determined whether the label carrier ends in the first end reading segment and the second end reading segment meet the preset label integrity condition.

[0046] ② Determine whether the first end reading segment and the second end reading segment meet the preset valid pairing conditions; if they do, then classify the relevant reading segment into the complete tag reading segment set; otherwise, classify the relevant reading segment into the incomplete tag reading segment set.

[0047] In this embodiment, the preset tag integrity conditions include at least one of the following conditions: the molecular tag sequence is located at a preset position of the read segment; the length of the molecular tag sequence is consistent with the preset length; there is a preset relative positional relationship between the molecular tag sequence and the target-specific primer region; the base quality of the molecular tag sequence meets the preset quality conditions; the number or proportion of undetermined bases in the molecular tag sequence is lower than a preset threshold; and the orientation or tag type of the molecular tag sequence is consistent with the preset molecular tag structure.

[0048] In this embodiment, the structures of the primer-related region molecular tag and the amplification product are as follows: Figure 2 As shown, the molecular tag is located at the 5′ end of the tag-bearing read segment and precedes the target-specific primer sequence. The preset length of the molecular tag is determined in advance during the amplification primer and molecular tag structure design stage. For any given amplification primer or detection scheme, the molecular tag adopts a preset fixed length; the preset length of the molecular tag corresponding to different amplification primers or detection schemes can be different, and the preset length can be selected in the range of 6 to 12 bases.

[0049] When performing molecular tag parsing, the starting position of the target-specific primer in the read segment is first determined by target-specific primer sequence matching. Then, according to the preset tag length corresponding to the amplification primer, a corresponding number of continuous bases are extracted from the starting position of the target-specific primer to the 5′ end of the read segment as the molecular tag sequence.

[0050] Specifically, when the molecular tag corresponding to a certain amplification primer has a preset length of L bases, where L is a predetermined integer from 6 to 12, L bases are continuously extracted from the starting position of the target-specific primer as the molecular tag sequence. When the length of the extracted molecular tag sequence is equal to the preset length L, the molecular tag is located before the target-specific primer sequence, and the molecular tag sequence meets the preset quality conditions, it is determined that the corresponding tag-bearing end meets the preset tag integrity conditions.

[0051] When a target-specific primer sequence is not identified, the number of extractable bases before the start position of the target-specific primer is less than the preset length L, the length of the extracted molecular tag sequence is inconsistent with the preset length L, or the molecular tag sequence does not meet the preset quality conditions, it is determined that the corresponding tag carrier end does not meet the preset tag integrity conditions.

[0052] In this embodiment, the preset valid pairing conditions include at least one of the following conditions: the first end read and the second end read have the same read name subject or pairing identifier; the first end read and the second end read have corresponding first end read types and second end read types; the first end read and the second end read are aligned to the same target region; the alignment direction of the first end read and the second end read conforms to the preset paired-end sequencing direction relationship; and the insert fragment length determined according to the alignment position of the first end read and the second end read is within a preset range.

[0053] Specifically, when the first end read and the second end read have the same read name or pairing identifier, and are respectively marked as the first end read and the second end read, and are aligned to the same target region, with the alignment directions being opposite and conforming to the preset paired sequencing direction relationship, and the length of the insert fragment determined based on the alignment positions of the two ends is 80 to 110 bases, then the first end read and the second end read are determined to meet the preset valid pairing conditions.

[0054] For the dual-tag design, it is determined whether the molecular tags in the first and second read segments meet the corresponding tag integrity conditions. In this embodiment, there is no fixed tag sequence combination correspondence between the first and second molecular tags. Therefore, whether the first and second molecular tags belong to a preset tag combination is not a necessary condition for determining whether the first and second read segments meet the preset valid pairing conditions.

[0055] It should be noted that the complete tag reading segment set in this embodiment refers to the set of related reading segments in which the tag-bearing ends of the molecular tag to be parsed according to the preset molecular tag structure all meet the preset tag integrity conditions, and the first end reading segment and the second end reading segment meet the preset valid pairing conditions.

[0056] The incomplete tag read set includes at least one of the following reads: missing tag reads, incomplete tag reads, abnormal tag length reads, low-quality tag reads, abnormal tag position reads, abnormal tag structure reads, failed pairing reads, and reads with undeterminable pairing relationships.

[0057] Since missing tags, incomplete tags, abnormal tag structures, or read segments whose pairing relationships do not meet the preset valid pairing conditions may introduce event-level statistical bias, the incomplete tag read segment set is not used to construct the molecular tag family participating in the statistics of the total number of determinate pairing tag families and the number of pairing candidate state families.

[0058] In this embodiment, the incomplete tag reading segment set can be used for at least one of tag integrity statistics, regional coverage statistics, background status verification, abnormal reading segment prompts, quality control, or auxiliary report output.

[0059] Step 2: Construct molecular tag families based on the complete tag reading set, perform consensus determination on each molecular tag family, and select determinate double-end paired tag families from the molecular tag families with clear consensus status.

[0060] (2.1) Based on the molecular tag sequence and fragment feature information in the complete tag read set, construct molecular tag families and determine the independent fragments in each molecular tag family.

[0061] In this embodiment, the segment feature information includes at least one of the following: target region information, segment comparison position information, segment start and end position information, segment direction information, and double-end pairing relationship information.

[0062] For any relevant read segment in the complete tag read set, its molecular tag sequence is obtained, and the molecular tag sequence is standardized according to a preset standardization rule.

[0063] For dual-label designs, the first and second molecular labels can be combined or standardized in a preset order to obtain combined molecular labels for family merging. The first and second molecular labels are not required to belong to a pre-defined fixed label combination.

[0064] The standardized molecular tag sequence is combined with the fragment feature information to obtain the molecular identifier of the corresponding related read segment; related read segments with the same molecular identifier are grouped into the same molecular tag family.

[0065] In one specific embodiment, the molecular identifier includes a standardized molecular tag sequence, the chromosome or target region where the read is located, the start position of the fragment, the end position of the fragment, and information on the pairing type or orientation of the two ends.

[0066] The start and end positions of the segment can be determined based on the comparison position of the first and second end reading segments and the length of the inserted segment. When the segment range cannot be determined based on complete end information, the corresponding segment features can be determined based on the available reading segment comparison positions, and the relevant reading segments can be marked or processed according to preset rules.

[0067] In another implementation, molecular tags that may differ due to amplification or sequencing errors can be corrected or merged based on the similarity between molecular tag sequences. The similarity can be represented by Hamming distance, edit distance, or other sequence distances, with the distance threshold determined based on the molecular tag length and a preset error tolerance range.

[0068] Within any molecular tag family, first and second end reads with the same read name or paired identifier are identified as the same independent segment. Each pair of corresponding first and second end reads is counted as an independent segment.

[0069] The number of independent fragments is used to characterize the family depth or amplification support depth of the corresponding molecular tag family.

[0070] (2.2) Based on the reference sequence information, determine the state supported by each independent fragment, and use the independent fragment as the statistical unit to make consensus judgment on each molecular tag family, determine the consensus state of each molecular tag family, and determine the molecular tag family with a clear consensus state.

[0071] The schematic diagram of molecular tag family consensus determination and family status classification in this embodiment is as follows: Figure 4 As shown.

[0072] In this embodiment, for any molecular tag family, the first and second reads of each independent fragment within the molecular tag family are compared or matched with reference sequence information. Based on the base sequence, insertion sequence, deletion sequence or other sequence characteristics at the target region or target site, the state supported by the first and second reads is determined.

[0073] The states include reference states, candidate states, and other states.

[0074] The reference state can be a reference sequence state, background state, preset standard state, or other sequence states used as a comparison benchmark; the candidate state can be a preset target sequence state to be analyzed, target site state, target region state, mutation state, or other states to be analyzed; other states are states other than the reference state and candidate states.

[0075] For any independent segment, if both the first end read segment and the second end read segment can determine the supported state, and both support the same state, then the state is determined to be the state supported by the independent segment.

[0076] When both the first and second end read segments can determine the supported states, but they support different states, the independent segment is marked as a conflicting independent segment.

[0077] When at least one of the first and second end read segments cannot determine the supported state, the independent segment is marked as an undeterminable independent segment according to the preset state determination rule, or the initial support state of the independent segment is determined based on the end that can determine the state.

[0078] Preferably, when performing event-level statistics with high requirements for bilateral consistency, conflicting independent segments and undecidable independent segments are not considered as bilaterally consistent independent segments supporting the reference state or candidate state.

[0079] The read segment is able to determine the state it supports, meaning that the alignment interval of the read segment covers the target region or target site, the sequence information at the target region or target site can be classified into one of the reference state, candidate state or other states, and the alignment quality of the corresponding read segment and the base quality at the target position meet the preset quality conditions.

[0080] When a read does not cover the target region or target site, the base at the target position is undetermined, the sequence information at the target position cannot determine the state it supports, the read alignment quality is insufficient, or the base quality at the target position is insufficient, the read is classified as an undetermined read.

[0081] For any molecular tag family, count the number of independent fragments supporting the reference state, candidate state, and other states, and determine the total number of independent fragments participating in the consensus decision.

[0082] The support ratio for a given state is the ratio of the number of independent fragments supporting that state to the total number of independent fragments participating in the consensus decision.

[0083] The total number of independent segments participating in consensus determination can be determined according to preset consensus rules. In one implementation, the independent segments participating in consensus determination include independent segments that can determine the supporting state; in another implementation, conflicting independent segments or undeterminable independent segments can also be included in the family state distribution as undetermined states.

[0084] The state with the highest support ratio is determined, and the support ratio of this state is compared with the preset consensus ratio threshold of the corresponding molecular tag family.

[0085] When the state with the highest support ratio is unique, and its support ratio is greater than or equal to the preset consensus ratio threshold, the state is determined as the consensus state of the molecular tag family, and the molecular tag family is determined as a molecular tag family with a clear consensus state.

[0086] When the support ratio of the state with the highest support ratio is lower than the preset consensus ratio threshold, or when there are two or more states with the same number of supports and both being the highest, the molecular tag family is determined as a family without consensus.

[0087] When there is no independent fragment within a molecular tag family that can be used to determine a reference state, candidate state, or other state, the molecular tag family can be classified as an undecidable family.

[0088] In this embodiment, the preset consensus ratio threshold can be dynamically set according to the number of independent fragments participating in the consensus determination within the molecular tag family.

[0089] Specifically, when the number of independent segments participating in the consensus determination is less than or equal to a first quantity threshold, a first consensus ratio threshold is used; when the number of independent segments is greater than the first quantity threshold and less than or equal to a second quantity threshold, a second consensus ratio threshold is used; and when the number of independent segments is greater than the second quantity threshold, a third consensus ratio threshold is used.

[0090] The first consensus ratio threshold, the second consensus ratio threshold, and the third consensus ratio threshold can be set based on family depth, sequencing quality, amplification error rate, and historical validation data.

[0091] In one specific implementation, the first quantity threshold is 3, and the second quantity threshold is 10; when the number of independent fragments does not exceed 3, the consensus ratio threshold is 50%; when the number of independent fragments is 4 to 10, the consensus ratio threshold is 70%; and when the number of independent fragments is greater than 10, the consensus ratio threshold is 80%.

[0092] When a molecular tag family has a clear consensus state, the molecular tag family is determined as a reference state family, a candidate state family, or another state family based on its consensus state.

[0093] Conflicting independent segments and undeterminable independent segments can be counted separately for quality control, anomaly alerts, or pending review marking.

[0094] In one optional implementation, when the number or proportion of conflicting independent segments within a molecular tag family reaches a preset conflict condition, the molecular tag family can be further marked as a conflict state family; when there are no independent segments with a definite state within a molecular tag family, the molecular tag family can be marked as an undeterminable family.

[0095] (2.3) Screen out the families of molecular tags that can be identified as paired at both ends from the families of molecular tags that have a clear consensus, and determine the total number of families of paired at both ends.

[0096] The schematic diagram of the screening of dual-end paired tag families in this embodiment is shown below. Figure 5 As shown.

[0097] In this embodiment, for any molecular tag family with a clear consensus, it is determined whether there are independent segments with identifiable paired ends within the molecular tag family.

[0098] For any independent segment, when the first end read segment and the second end read segment have the same read segment name body or pairing identifier, respectively corresponding to the first end read segment and the second end read segment, and both the first end read segment and the second end read segment can determine the supported state, and both support the same state, the independent segment is determined as a determinate double-end paired independent segment.

[0099] When the first end read segment and the second end read segment support different states, the independent segment is identified as a conflicting independent segment and is not considered as a definite independent segment with consistent states at both ends.

[0100] When at least one of the first and second end read segments cannot determine the state it supports, the independent segment is marked as an undecidable independent segment and is not considered a decidable independent segment with consistent states at both ends.

[0101] A molecular tag family with a clear consensus state is identified as a determinate paired tag family when it contains at least one determinate paired independent fragment.

[0102] In one implementation, the number of independent fragments within the molecular tag family may also be required to reach a preset family depth threshold. In one specific implementation, the preset family depth threshold is 2, meaning the molecular tag family includes at least two independent fragments.

[0103] After traversing each molecular tag family, the total number of identifiable paired tag families is counted and denoted as N1.

[0104] Families without consensus are not included in the N1 statistics. Conflict-state families and undecidable families, if individually labeled, are also not included in the N1 statistics, but can be retained for quality control, anomaly alerts, background status verification, or auxiliary report output.

[0105] Step 3: Select candidate state families with consistent pairing from the families of identifiable pairing tags, determine the state analysis results of the target region or target site based on the number of candidate state families with consistent pairing, and verify the state analysis results.

[0106] (3.1) Filter out candidate state families with consistent two-ends from the families of identifiable two-end paired labels and determine their number.

[0107] For any candidate state, filter the molecular tag families whose consensus state is the candidate state from the families of determinate paired tags, and determine whether there are independent segments of the consistent candidate state within the molecular tag families.

[0108] For any independent segment, if both the first end read segment and the second end read segment of the independent segment can determine the supported state, and both the first end read segment and the second end read segment support the same candidate state, then the independent segment is determined as the independent segment of the two-end consistent candidate state corresponding to the candidate state.

[0109] When the consensus state of a definite pairwise paired tag family is the candidate state, and there is at least one independent segment of the pairwise consistent candidate state that supports the same candidate state within the molecular tag family, the molecular tag family is determined as the pairwise consistent candidate state family corresponding to the candidate state.

[0110] When the consensus state of a molecular tag family is a candidate state, but there is no independent segment within the molecular tag family where both the first end read and the second end read support the candidate state, the molecular tag family is not determined as a two-end consistent candidate state family.

[0111] After traversing each family of determinate paired labels, count the number of families of consistent paired candidate states corresponding to the candidate states, and denote it as N2.

[0112] The total number of identifiable paired tag families is denoted as N1. When multiple candidate states exist, the above filtering process is performed for each candidate state, and N2 is determined for each candidate state.

[0113] (3.2) Based on the number of consistent candidate state families at both ends, determine the state analysis results of the target region or target site.

[0114] In one implementation, the number N2 of the two-end consistent candidate state families is compared with a preset number threshold.

[0115] When N2 is greater than or equal to a preset quantity threshold, the quantity support condition for the target region or target site to reach the corresponding candidate state is determined; when N2 is less than the preset quantity threshold, the quantity support condition for the target region or target site not to reach the corresponding candidate state is determined.

[0116] The preset quantity threshold can be determined based on sequencing depth, amplification error rate, sequencing error rate, negative sample detection results, repeatability test results, or historical validation data.

[0117] In one specific implementation, the preset quantity threshold can be selected within the range of 2 to 5, for example, set to 3. By setting a minimum quantity threshold, the impact of single amplification errors, sequencing errors, or isolated anomalous events on the candidate state analysis results can be reduced, thereby improving the reliability of the candidate state analysis results.

[0118] In another implementation, the pairwise consistency amplification event support index I is calculated based on the total number N1 of determinate pairwise paired tag families and the number N2 of pairwise consistent candidate state families. A schematic diagram of the pairwise consistent candidate state family screening and pairwise consistency amplification event support index calculation in this embodiment is shown below. Figure 6 As shown.

[0119] When N1 is greater than 0, the support index I for paired-end consensus amplification events is calculated according to the following relationship: I = N2 / N1.

[0120] Where N1 is the total number of identifiable paired label families corresponding to the target region or target site, and N2 is the number of paired consistent candidate state families corresponding to any candidate state.

[0121] The support index I for the paired-end consistent amplification event is compared with a preset support index threshold. When I is greater than or equal to the preset support index threshold, the index support condition for the target region or target site to reach the corresponding candidate state is determined; when I is less than the preset support index threshold, the index support condition for the target region or target site not to reach the corresponding candidate state is determined.

[0122] When N1 is 0, the support index I for the two-end consistent amplification event is not calculated, and I is marked as "not applicable". At the same time, a prompt is output indicating that the number of effective families is insufficient or the state is undeterminable. This situation is not interpreted as the candidate state not existing.

[0123] When N1 is greater than 0, the numerical range of the paired-end consistent amplification event support index I is [0,1]. When outputting the analysis results, the paired-end consistent amplification event support index I can be output as a proportion, percentage, or fraction of ten thousand. If I is normalized, scaled, or corrected for quality weights, the processed value will be separately identified as a derived index. Candidate state support determination is based on the original support index calculated according to I=N2 / N1 and its corresponding threshold, or the threshold can be converted to the same scale.

[0124] The preset support index threshold is located in the range of [0,1], and the preset support index threshold can be determined based on the support index distribution of negative samples or normal background samples.

[0125] In one specific implementation, under the same experimental and analytical conditions, multiple negative samples or normal background samples without corresponding candidate states are sequenced and analyzed. The support index of each sample at the target region or target site is calculated, and the sum of the average value of the support index and three times the standard deviation is used as the preset support index threshold.

[0126] In other implementations, a preset percentile of the background support index, the upper limit of the confidence interval, or a verified fixed threshold may be used as the preset support index threshold.

[0127] The state analysis results of the target region or target site can be determined based on at least one of N2, I and its corresponding preset threshold.

[0128] In the implementation method that uses a quantity threshold for determination, the corresponding candidate state is determined to meet the quantity support condition based on whether N2 reaches a preset quantity threshold.

[0129] In the implementation method that uses a support index threshold for determination, the corresponding candidate state is determined to meet the index support condition based on whether I reaches the preset support index threshold.

[0130] In the implementation using joint determination, when both N2 and I meet their respective preset support conditions, the analysis result indicating that the candidate state support condition has been met is output; when at least one of N2 and I does not meet the corresponding preset support condition, the analysis result indicating that the candidate state support condition has not been met is output; when the preset manual review trigger condition is met, a prompt indicating that the review is pending can also be output.

[0131] In this embodiment, four sets of amplification and sequencing data containing molecular tags, A, B, C, and D, are used as examples. The calculation results of their paired-end consistent amplification event support index and the state analysis results of the target region are shown in Table 1.

[0132] Table 1. Calculation results of the paired-end congruent amplification event support index for molecularly tagged amplification sequencing data. The data listed in Table 1 are example data used to illustrate the calculation of the paired-end consistent amplification event support index and the process of candidate state determination, and do not represent actual experimental data.

[0133] It should be noted that in Table 1, the support index I for two-end consistent amplification events is calculated according to I = N2 / N1 and output as a percentage. Table 1 uses a joint determination method of N2 and I. Only when both the number of two-end consistent candidate state families N2 and the support index I for two-end consistent amplification events reach their respective preset thresholds is the corresponding candidate state determined to have met the preset support condition. Although the N2 of record B reaches the example number threshold of 3, its support index I has not reached the preset support index threshold, therefore it is determined that the preset support condition has not been met.

[0134] In this example, the preset quantity threshold is 3, and the preset support index threshold is 0.01%.

[0135] (3.3) Calculate the high confidence candidate state support index and verify the state analysis results of the target region or target site.

[0136] When the target region or target site reaches the preset support conditions of the corresponding candidate state according to step (3.2), the high confidence review process is initiated, and the support index J and relative deviation degree D of the high confidence candidate state are calculated according to the following method.

[0137] When the target region or target site does not meet the preset support conditions of the corresponding candidate state, the high confidence review process is not initiated, the high confidence candidate state support index J and the relative deviation degree D are not calculated, and J and D are marked as "not applicable". At the same time, the analysis results that do not meet the support conditions of the candidate state are output.

[0138] From the families of identifiable paired tags, further screen molecular tag families that meet the preset high confidence conditions, identify them as high confidence tag families, and count the total number N3 of high confidence tag families.

[0139] In this embodiment, the high-confidence tag family satisfies the following preset high-confidence condition: The molecular tag family has a clear consensus. The number of independent fragments in the molecular tag family reaches a preset family depth threshold; The molecular tag family corresponds to the same target region or target site; The molecular tag family includes at least one first end read segment and a second end read segment that can both determine the supported state, and both the first end read segment and the second end read segment support independent fragments of the consensus state of the molecular tag family. The molecular tag family does not belong to the family without consensus, the family in conflict, or the family that cannot be determined. The molecular tag family meets preset alignment quality, base quality, or other quality control conditions.

[0140] The preset family depth threshold can be determined based on sequencing depth, amplification efficiency, background error level, and historical validation data.

[0141] In one specific implementation, the preset family depth threshold is set to 2, meaning that a high-confidence label family includes at least two independent segments.

[0142] For any candidate state, a consensus state is selected from the high-confidence tag family as the candidate state, and a molecular tag family that contains at least one independent fragment whose first end read and second end read both support the candidate state is determined as the high-confidence candidate state family corresponding to the candidate state.

[0143] The number of the high-confidence candidate state families is counted and denoted as N4.

[0144] When N3 is greater than 0, the support index J for high-confidence candidate states is calculated according to the following relationship: J = N4 / N3.

[0145] Where N3 is the total number of high-confidence label families, and N4 is the number of high-confidence candidate state families corresponding to the candidate states.

[0146] When N3 is 0, the high-confidence candidate state support index J is not calculated, and J is marked as "not applicable". At the same time, a prompt is output indicating that the number of high-confidence label families is insufficient or that high-confidence review cannot be completed.

[0147] The high-confidence candidate state support index J has a numerical range of [0,1]. When outputting the analysis results, J can be output as a proportion, percentage, or fraction of ten thousand. If J is normalized, scaled, or corrected for quality weights, the processed value will be identified separately as a derived index; the consistency verification of I and J should use I and J under the same numerical scale.

[0148] The high-confidence candidate state support index J is used to verify the candidate state support reflected by the paired-end consistent amplification event support index I.

[0149] When a high-confidence review process is initiated, and N3 is greater than 0 and at least one of I and J is greater than 0, the relative deviation degree D is calculated according to the following relationship: D=|IJ| / max(I,J).

[0150] When the relative deviation degree D is less than or equal to the preset verification threshold, it is determined that the support index I of the two-end consistent amplification event is consistent with the candidate state support result reflected by the high confidence candidate state support index J.

[0151] When the relative deviation degree D is greater than the preset review threshold, the system outputs a support index inconsistency prompt, an abnormal prompt, or a pending review mark to prompt further verification.

[0152] The verification is used to evaluate whether the proportion of candidate state support reflected by molecular tag families of different quality levels is consistent, and is not used to prove that the state analysis results of the target region or target site are absolutely correct.

[0153] In one implementation, the preset review threshold can be selected within the range of 20% to 30%. The specific value of the preset review threshold can be determined based on negative samples, positive samples, repeatability test data, statistical fluctuation range at different family depths, or historical validation data.

[0154] When the target region or target site does not meet the candidate state support conditions, the relative deviation degree calculation is not performed, and the relative deviation degree D is marked as "not applicable".

[0155] When N3 is 0, the high-confidence candidate state support index J and relative deviation degree D are not calculated, and J and D are marked as "not applicable". At the same time, a prompt indicating insufficient high-confidence verification data is output.

[0156] In this embodiment, taking four sets of molecular tag-containing amplification sequencing data A, B, C and D as examples, the verification results of their paired-end consistent amplification event support index I and high confidence candidate state support index J are shown in Table 2.

[0157] Table 2. Verification results of the support index for paired-end consistent amplification events and the support index for high-confidence candidate states. In Table 2, the high-confidence candidate state support index J is used to further verify the analysis results that have met the candidate state support conditions. When the target region or target site meets the candidate state support conditions and N3 is greater than 0, the high-confidence candidate state support index J is calculated according to J=N4 / N3, and the relative deviation degree D is calculated according to D=|IJ| / max(I,J). When the target region or target site does not meet the candidate state support conditions, or N3 is 0, high-confidence verification is not initiated, J and D are not calculated, and J and D are marked as "not applicable," with a corresponding prompt output. The data listed in Table 2 are example data used to illustrate the paired-end consistent amplification event support index, the high-confidence candidate state support index, and the verification process, and do not represent actual experimental data.

[0158] This invention screens candidate families of paired-end matching from families of determinate paired-end matching tags. It calculates the paired-end matching amplification event support index I based on the ratio of the number of candidate families of paired-end matching to the total number N1 of determinate paired-end matching tag families. This index can be combined with the absolute number of N2 and the relative proportion of I for joint determination, thus considering both the event quantity support and the proportion support of the candidate states. Furthermore, this invention can also verify the state analysis results using the high-confidence candidate state support index J and the relative deviation D between I and J, improving the accuracy, statistical consistency, and verifiability of amplification sequencing data analysis.

[0159] (3.4) Calculate the reference state support index and summarize the candidate state support index and quality control indicators.

[0160] For any target region or target site, select molecular tag families with a consensus state as the reference state from the families of identifiable paired tags, and determine whether there are independent segments with a consistent reference state at both ends within the molecular tag families.

[0161] For any independent segment, if both the first and second end read segments of the independent segment can determine the supported state, and both the first and second end read segments support the reference state, then the independent segment is determined as a dual-end consistent reference state independent segment.

[0162] When a consensus state of a identifiable paired tag family is a reference state, and there is at least one independent segment of a consistent reference state within the molecular tag family, the molecular tag family is identified as a consistent reference state family.

[0163] After traversing each family of determinate paired labels, count the number of families of consistent reference states and denote it as N0.

[0164] When the total number N1 of identifiable paired tag families is greater than 0, the reference state support index R is calculated according to the following relationship: R = N0 / N1.

[0165] Where N0 is the number of families with consistent reference states at both ends, and N1 is the total number of families with identifiable paired tags at both ends.

[0166] In step (3.2), the number of families of two-end consensus candidate states N2 has been determined for any candidate state, and the support index I of the two-end consensus amplification event corresponding to the candidate state is calculated according to I=N2 / N1.

[0167] When multiple candidate states exist, determine the number N2 of the two-ends-consistent candidate state families corresponding to each candidate state, and calculate the two-ends-consistent amplification event support index I corresponding to each candidate state.

[0168] Since the number of two-end consistent reference state families N0 and the number of two-end consistent candidate state families corresponding to each candidate state N2 only count molecular tag families that simultaneously meet the corresponding consensus state and two-end consistent support conditions, and the total number of two-end paired tag families N1 can be determined to include other state families or families that do not meet the corresponding two-end consistent support conditions, the sum of the reference state support index R and the two-end consistent amplification event support index I corresponding to each candidate state is not greater than 1, and can be less than 1 in the above case.

[0169] When N1 is 0, the reference state support index R and the paired-end consistent amplification event support index I corresponding to each candidate state are not calculated, and R and I are marked as "not applicable". At the same time, a prompt is output indicating that the number of effective families is insufficient or the status of the target region or target site is undetermined. This situation is not interpreted as the absence of a reference state or candidate state.

[0170] In this embodiment, quality control indicators are also calculated based on the segment stratification results, independent fragment determination results, molecular tag family consensus determination results, and state analysis results.

[0171] The quality control indicators include at least one of the following: The proportion of complete tag reads to the total number of reads related to the target region or target site; The proportion of incomplete tag reads to the total number of reads related to the target region or target site; The proportion of families without consensus to the total number of molecular tag families constructed based on complete tag read sets; The proportion of conflicting independent fragments to the total number of independent fragments involved in the family consensus determination; The proportion of undeterminable independent fragments to the total number of independent fragments involved in the family consensus determination; The proportion of conflict-state families to the total number of molecular tag families constructed based on the complete tag read set; The proportion of undeterminable families to the total number of molecular tag families constructed based on the complete tag read set; Number of families of consistent reference states with both ends: N0; The total number of paired tag families, N1, can be determined. The number of bi-end consistent candidate state families determined for each candidate state is N2; The total number of high-confidence label families is N3; The number of high-confidence candidate state families determined for each candidate state is N4; Reference state support index R; The support index I for the two-end consistent amplification event is determined for each candidate state; The support index J for high-confidence candidate states is determined for each candidate state; And the relative deviation D between the two-end consistent amplification event support index I and the high-confidence candidate state support index J determined when initiating high-confidence review.

[0172] When the target region or target site reaches the preset support conditions of the corresponding candidate state, initiates high confidence verification and N3 is greater than 0, the relative deviation degree D is determined according to the method described in step (3.3).

[0173] When the target region or target site does not meet the preset support conditions for the corresponding candidate state, high confidence verification is not initiated, or N3 is 0, the relative deviation degree D is not calculated, and the relative deviation degree D is marked as "not applicable". At the same time, a prompt is output according to the specific situation that the candidate state support conditions are not met or the high confidence verification data is insufficient.

[0174] For quality control indicators that need to be expressed in proportional form, when the corresponding denominator is greater than 0, the quality control indicator is calculated according to the ratio of the corresponding numerator to the denominator; when the corresponding denominator is 0, the quality control indicator is not calculated, and it is marked as "inapplicable" or "insufficient data" according to the reason why the denominator is 0, and a corresponding prompt is output. The situations where the denominator is 0 include at least one of the following: the total number of reads related to the target region or target site is 0, the total number of molecular tag families constructed based on the complete tag read set is 0, or the total number of independent fragments participating in the family consensus determination is 0.

[0175] The quality control indicators are compared with their respective preset quality control thresholds.

[0176] In one implementation, when the total number N1 of the determinate paired tag families reaches a preset minimum number, the proportion of complete tag reading segments reaches a preset lower limit, and all the quality control indicators that are preset to be met among the proportion of families without consensus, proportion of conflicting independent segments, proportion of undeterminable independent segments, proportion of conflicting families, and proportion of undeterminable families do not exceed their respective preset upper limits, the target region or target site is determined to meet the corresponding quality control conditions.

[0177] When N1 is lower than the preset minimum quantity, the proportion of complete tag reading segments is lower than the preset lower limit, or any quality control indicator that is preset to be met fails to meet the corresponding quality control conditions, the output will be prompted as insufficient valid data, quality control failure, insufficient reliability of results, or pending review.

[0178] The preset quality control threshold can be determined based on the sequencing platform, amplification system, molecular tag structure, target region or target site characteristics, sequencing depth, negative sample test results, positive sample test results, repeatability test results, or historical validation data.

[0179] In this embodiment, the final output includes at least one of the following: Target area or target location information; Reference status information and candidate status information; Number of complete tag segments read and number of incomplete tag segments read; Total number of molecular tag families; Number of families without consensus, number of families in conflict, and number of families whose outcomes are undecidable; Number of families of consistent reference states with both ends: N0; The total number of paired tag families, N1, can be determined. The number of bi-terminal consistent candidate state families corresponding to each candidate state is N2; The total number of high-confidence label families is N3; The number of high-confidence candidate state families corresponding to each candidate state is N4; The reference state supports the index R or its inapplicable label; The double-ended consistent amplification event support index I or its inapplicable label corresponding to each candidate state; The support index J of the high-confidence candidate state corresponding to each candidate state or its inapplicable label; The degree of relative deviation D or its inapplicable marking; Candidate states support the judgment result; Quality control results; Error message; Insufficient data for high-confidence review; And marked as pending review.

[0180] Device Example: This invention also provides a molecular tag-based amplification event-level sequencing data analysis device, the schematic diagram of which is shown below. Figure 7 As shown.

[0181] The device includes an input module, a core analysis module, an output and interaction module, and a data storage and management module.

[0182] The core analysis module includes a segment reading processing module, a family analysis module, and a status analysis and verification module.

[0183] The input module is used to acquire amplified sequencing data containing molecular tags and preset sequence structure reference information, and send the amplified sequencing data and preset sequence structure reference information to the read processing module.

[0184] The molecularly tagged amplification and sequencing data refers to the data obtained by sequencing the molecularly tagged amplification products. The data format of the amplification and sequencing data includes at least one of the following: FASTQ format, SAM format, BAM format, CRAM format, or other sequencing data formats.

[0185] The amplification product carries a resolvable molecular tag. The molecular tag can be introduced into the amplification product through amplification primers, primer extension regions, primer-related adapter regions, amplification product end regions, or combinations thereof, and can be resolved from a preset tag-bearing end or preset position of the sequencing read.

[0186] The preset sequence structure reference information includes at least one of the following: target region information, target site information, reference sequence information, primer structure information, and molecular tag structure information.

[0187] The reference sequence information includes a reference genome sequence, reference gene sequence, or preset standard sequence corresponding to the target region or target site. It is used to compare or match sequencing reads, determine the corresponding position of the sequencing reads on the reference sequence, and select reads related to the target region or target site accordingly.

[0188] The read segment processing module is used to acquire relevant read segments of the target region or target site, parse the molecular tag sequence in the relevant read segments, determine at least one of the molecular tag's position, length, sequence and quality, determine whether the corresponding tag-bearing end meets the preset tag integrity condition, and determine whether the first end read segment and the second end read segment meet the preset valid pairing condition, so as to divide the relevant read segments into a complete tag read segment set and an incomplete tag read segment set.

[0189] The incomplete tag read set is not used to construct the molecular tag family for the statistics of the total number of determinate paired tag families N1, the number of paired consistent candidate state families N2, the total number of high-confidence tag families N3, and the number of high-confidence candidate state families N4.

[0190] The family analysis module is used to construct molecular tag families based on the complete tag read set, and to identify the first and second end reads with the same read name or paired identifier as the same independent fragment.

[0191] The family analysis module is also used to determine the state supported by each independent segment, and to make consensus judgments on each molecular tag family using independent segments as statistical units, so as to obtain at least one of the following: molecular tag family with clear consensus state, family without consensus, family with conflict state, and undecidable family.

[0192] The family analysis module is also used to screen determinate double-paired tag families from molecular tag families with a clear consensus state, which meet the double-end pairing condition and the state determinate condition, and to determine the total number N1 of the determinate double-end paired tag families.

[0193] The state analysis and verification module is used to filter out the two-end consistent candidate state families from the families of identifiable two-end paired tags for any candidate state, and determine the number N2 of the corresponding two-end consistent candidate state families.

[0194] The state analysis and verification module is also used to calculate the double-end consistent amplification event support index I based on the quantitative relationship between N2 and N1, and to determine whether the target region or target site has reached the support conditions of the corresponding candidate state based on at least one of N2, I, preset quantity threshold and preset support index threshold.

[0195] The state analysis and verification module is also used to further filter high-confidence label families that meet preset high-confidence conditions from the determinate double-ended paired label families, and to filter high-confidence candidate state families from the high-confidence label families, respectively determining the total number N3 of high-confidence label families and the number N4 of high-confidence candidate state families corresponding to each candidate state.

[0196] The state analysis and verification module is also used to calculate the high-confidence candidate state support index J based on the quantitative relationship between N4 and N3, and to verify the relative deviation degree D between the double-ended consistent amplification event support index I and the high-confidence candidate state support index J when the target region or target site reaches the corresponding candidate state support condition.

[0197] The state analysis and verification module is also used to filter out double-end consistent reference state families from the families of identifiable double-end paired tags, determine the number N0 of double-end consistent reference state families, and calculate the reference state support index R based on the quantitative relationship between N0 and N1.

[0198] The status analysis and verification module is also used to determine at least one quality control indicator, and to determine the quality control result based on the comparison result between the quality control indicator and the corresponding preset quality control threshold.

[0199] The output and interaction module is used to output at least one of the following: state analysis results of the target region or target site, reference state information, candidate state information, N0, N1, N2, N3, N4, reference state support index R, paired-end consistent amplification event support index I, high confidence candidate state support index J, relative deviation degree D, quality control results, abnormal prompts, and markers to be reviewed.

[0200] The output and interaction module is also used to receive analysis parameters or operation instructions input by the user, display the device's operating status, and retrieve and display analysis results, quality control results, operation logs, or abnormal information.

[0201] The data storage and management module is used to store at least one of the following: input amplification and sequencing data, preset sequence structure reference information, intermediate data generated during the analysis process, status analysis results, quality control results, configuration parameters, judgment rules, preset thresholds, device operation logs, and audit records.

[0202] The intermediate data includes at least one of the following: numerical values ​​of N0, N1, N2, N3, and N4; statistical results of various molecular tag families; independent fragment status results; reference status support index R; paired-end consistent amplification event support index I; high-confidence candidate status support index J; and relative deviation degree D.

[0203] The judgment rules and preset thresholds include at least one of the following: preset tag integrity condition, preset valid pairing condition, molecular tag family merging condition, state determination condition, consensus ratio threshold, family depth threshold, candidate state quantity threshold, support index threshold, verification threshold, and quality control threshold.

[0204] The input module, segment processing module, family analysis module, status analysis and verification module, output and interaction module, and data storage and management module can be implemented through software, hardware, or a combination of software and hardware.

[0205] Electronic device example: The present invention also provides an electronic device, the schematic diagram of which is shown below. Figure 8 As shown.

[0206] The electronic device includes a processor and a memory, with the processor communicatively connected to the memory. The electronic device may also include at least one of an input device, an output device, a communication interface, and a storage device, with each component connected via a bus or other communication method.

[0207] The memory stores a computer program, and the processor executes the computer program to implement the molecular tag-based amplification event-level sequencing data analysis method described in the embodiments of the present invention.

[0208] The input device is used to receive amplification and sequencing data containing molecular tags, preset sequence structure reference information, analysis parameters, or user operation instructions.

[0209] The output device is used to display or output the status analysis results, reference status information, candidate status information, support index, quality control results, abnormal prompts, or pending verification marks of the target area or target site.

[0210] The storage device is used to store at least one of the following: raw amplification and sequencing data, reference information, intermediate analysis data, status analysis results, quality control results, configuration parameters, and operation logs.

[0211] The communication interface is used to enable data transmission between the electronic device and sequencing equipment, data server, storage device, user terminal or other electronic devices.

[0212] The electronic device may be a server, computer workstation, personal computer, cloud computing device, or other electronic device capable of executing the computer program.

[0213] Storage medium examples: The present invention also provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it implements the molecular tag-based amplification event-level sequencing data analysis method described in the embodiments of the present invention.

[0214] The computer-readable storage medium includes at least one of read-only memory, random access memory, solid-state memory, magnetic storage medium, optical storage medium, or other computer-readable storage medium capable of storing computer programs.

Claims

1. A method for analyzing amplification event-level sequencing data based on molecular tags, wherein the amplification sequencing data is obtained by sequencing amplification products carrying resolvable molecular tags, characterized in that, Includes the following steps: (1) Determine each relevant read of the target region or target site in the amplified sequencing data and the molecular tag sequence in each relevant read, wherein each relevant read consists of a first end read and a second end read with a corresponding relationship; based on the molecular tag sequence, screen complete tag reads from each relevant read; (2) Construct molecular tag families based on each complete tag reading segment; within each molecular tag family, determine the first end reading segment and the second end reading segment with corresponding relationship as the same independent segment; use the independent segment as the statistical unit to perform family consensus determination on each molecular tag family, determine the consensus state of each molecular tag family, and screen out the determinable double-end paired tag families from the molecular tag families with clear consensus state. (3) Filter out candidate states with consistent pairing from the families of identifiable paired tags and determine the number of such candidate states with consistent pairing. The state analysis results of the target region or target site are determined based on the number of the two-end consistent candidate state families.

2. The method for analyzing molecular tag-based amplification event-level sequencing data according to claim 1, characterized in that, Methods for selecting complete tagged segments from various related segments include: Based on the preset molecular tag structure, the tag-bearing end in the first and second end of each relevant read segment is determined, and the comprehensive attribute information of the molecular tag of the tag-bearing end is obtained. For any set of related reading segments with corresponding relationships, based on the comprehensive attribute information of the molecular tag, it is determined whether the tag carrier end of the molecular tag that needs to be parsed according to the preset molecular tag structure meets the preset tag integrity condition. If all the tag carrier ends meet the preset tag integrity condition, then determine whether the first end reading segment and the second end reading segment meet the preset valid pairing condition; if they do, then divide the relevant reading segments into complete tag reading segments.

3. The method for analyzing molecular tag-based amplification event-level sequencing data according to claim 1, characterized in that, Methods for constructing molecular tag families include: The molecular tag sequence and fragment feature information of each complete tag read are combined to obtain the molecular identifier of each complete tag read. Complete tag reads with the same molecular identifier are grouped into the same molecular tag family.

4. The method for analyzing molecular tag-based amplification event-level sequencing data according to claim 1, characterized in that, Methods for identifying families of paired tags include: For any molecular tag family with a clear consensus, determine whether each independent segment is a determinate double-end paired independent segment. If at least one independent segment is a determinate double-end paired independent segment, then the molecular tag family is determined as a determinate double-end paired tag family.

5. The method for analyzing molecular tag-based amplification event-level sequencing data according to claim 4, characterized in that, Methods for determining whether each independent segment is a determinate paired independent segment include: For any independent segment, determine whether the first end read segment and the second end read segment of the independent segment can determine the supported state; if both the first end read segment and the second end read segment can determine the supported state, and the first end read segment and the second end read segment support the same state, then the independent segment is determined to be a pairwise independent segment.

6. The method for analyzing molecular tag-based amplification event-level sequencing data according to claim 5, characterized in that, Methods for selecting candidate state families with consistent outcomes at both ends include: For any determinate dual-end paired tag family, determine whether the consensus state of the determinate dual-end paired tag family is a candidate state; When the consensus state of the determinate double-end pairing tag family is a candidate state, determine whether the first end read segment and the second end read segment of each determinate double-end pairing independent segment in the determinate double-end pairing tag family both support the same candidate state. If at least one determinates a pair of independent segments, both the first and second end read segments support the same candidate state, then the determinate pair of paired tags family is determined as a pairwise consistent candidate state family.

7. The method for analyzing event-level sequencing data based on molecular tags according to claim 1, characterized in that, Methods for determining the state analysis results of target regions or target sites include: Determine the total number of determinate paired tag families and the number of paired consistent candidate state families; The support index for double-end consistent amplification events is calculated based on the total number of determinate double-end paired tag families and the number of double-end consistent candidate state families. Based on the dual-end consistent amplification event support index, it is determined whether the target region or target site meets the support conditions for the corresponding candidate state.

8. The method for analyzing molecular tag-based amplification event-level sequencing data according to claim 7, characterized in that, The method for determining whether the target region or target site meets the support conditions of the corresponding candidate state includes: comparing the number of two-end consistent candidate state families with a preset number threshold, and comparing the two-end consistent amplification event support index with a preset support index threshold; when the number of two-end consistent candidate state families is greater than or equal to the preset number threshold, and the two-end consistent amplification event support index is greater than or equal to the preset support index threshold, the target region or target site is determined to meet the support conditions of the corresponding candidate state.

9. The method for analyzing molecular tag-based amplification event-level sequencing data according to claim 7, characterized in that, Methods for determining the state analysis results of target regions or target sites also include: From the families of identifiable paired labels, select high-confidence label families and high-confidence candidate state families, and determine the total number of high-confidence label families and the number of high-confidence candidate state families, respectively. The high-confidence candidate state support index is calculated based on the total number of high-confidence label families and the number of high-confidence candidate state families. The state analysis results were reviewed based on the high-confidence candidate state support index.

10. The method for analyzing molecular tag-based amplification event-level sequencing data according to claim 9, characterized in that, Methods for verifying the results of the status analysis include: Based on the two-end consistent amplification event support index and the high-confidence candidate state support index, the relative deviation between the two is calculated. Based on the degree of relative deviation, determine whether the state analysis results are correct.

11. A molecular tag-based amplification event-level sequencing data analysis device, characterized in that, It includes an input module and a core analysis module; The core analysis module is used to obtain analysis results based on the received data using the molecular tag-based amplification event-level sequencing data analysis method as described in any one of claims 1-10.

12. An electronic device, characterized in that, It includes a processor and a memory, wherein the memory stores a computer program, and when the processor executes the computer program, it implements the molecular tag-based amplification event-level sequencing data analysis method as described in any one of claims 1 to 10.

13. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the molecular tag-based amplification event-level sequencing data analysis method as described in any one of claims 1 to 10.

Images