An E3 ubiquitin ligase encoding gene recognition method, device and computer equipment

Abstract

The application provides an E3 ubiquitin ligase coding gene identification method, device and computer equipment, and belongs to the field of biological information analysis. The method comprises the following steps: obtaining a proteome sequence file and a protein-transcript-gene mapping file of a target species; determining a candidate protein set; obtaining domain annotation results from multiple domain database sources; mapping the multiple domain annotation results to the same protein sequence coordinate system, judging whether the domain annotations from different sources correspond to the same core domain; fusing the multiple source annotations of the same core domain; determining a core domain integrity index, judging the candidate proteins, and generating an evidence state record; merging multiple candidate proteins belonging to the same gene identifier to the gene level to obtain an E3 ubiquitin ligase coding gene directory of the target species in the form of a basic unit of a gene identifier, so that the judgment fluctuation caused by different database sources is eliminated.

Description

Technical Field

[0001] This invention belongs to the field of bioinformatics analysis, specifically relating to a method, apparatus, and computer device for identifying E3 ubiquitin ligase-encoded genes. Background Technology

[0002] E3 ubiquitin ligases are key regulators in the ubiquitin-proteasome system that determine substrate specificity, and are widely involved in biological processes such as plant and animal development, stress response, immune regulation, signal transduction, and protein homeostasis. The E3 family mainly includes types such as U-box / PUB, HECT, RING, and RBR, with different types having different core domains, accessory domain combinations, and boundary member characteristics.

[0003] When identifying E3 ubiquitin ligase-encoding genes at the genome level, existing methods typically rely on single HMM domain searches, keyword searches, or manual screening, requiring the processing of heterogeneous data from multiple sources, including whole proteome sequences, genome annotation files, domain models, reference protein sequences, and domain annotation results from multiple databases. However, due to the diversity of E3 family domain types and the short sequences of some domains, annotation results for the same protein from different domain databases (such as Pfam, SMART, PROSITE, and InterPro) often differ in domain coordinate definition and type determination, leading to inconsistent judgments based on different database sources. Furthermore, multiple transcripts or protein variants of the same gene can easily lead to duplicate counting; boundary member domains such as RING-like, NSMCE2-like, and RBBP6-like exhibit atypical characteristics and are easily misclassified as E3 family members. Existing methods lack a mechanism to unify and analyze domain annotation results from multiple databases under the same coordinate system, resulting in unstable identification results for E3 ubiquitin ligase-encoding genes, inconsistent gene hierarchy directories, high manual verification costs, and difficulty in reproducing the results during cross-species batch processing. Summary of the Invention

[0004] To address the problem of inconsistent annotation results from different domain database sources in the prior art, which leads to unstable identification results of E3 ubiquitin ligase-encoded genes in terms of domain coordinates and type determination, this invention provides a method, apparatus, and computer device for identifying E3 ubiquitin ligase-encoded genes.

[0005] To achieve the above objectives, the present invention provides the following technical solution: A method for recognizing E3 ubiquitin ligase-encoded genes, the method comprising: Obtain the proteome sequence file and protein-transcript-gene mapping file of the target species; Based on the domain model corresponding to E3 ubiquitin ligase, HMM domain search was performed on the proteome sequence to obtain a set of candidate proteins. For each candidate protein in the candidate protein set, domain annotation results are obtained from multiple domain database sources. The annotation results include domain type, start coordinates, end coordinates, and source. Domain annotation results from multiple databases are mapped to the protein sequence coordinate system of the same candidate protein. Domain intervals from different sources are sorted using the protein identifier as an index. The overlap and coordinate offset between domain intervals from different sources are calculated. Based on the overlap and coordinate offset, it is determined whether domain annotations from different sources correspond to the same core domain. For multiple annotations determined to correspond to the same core domain, they are fused to generate fused domain coordinates and the number of fused sources is recorded. Based on the fused domain coordinates, a core domain integrity index is determined. Based on the core domain integrity index, candidate proteins are determined, and evidence state records containing boundary state labels are generated. Based on the protein-transcript-gene mapping file and evidence status record, multiple candidate proteins belonging to the same gene identifier are merged into the gene level to obtain the E3 ubiquitin ligase-encoded gene catalog of the target species with gene identifier as the basic unit.

[0006] Optionally, the step of merging multiple candidate proteins belonging to the same gene identifier to the gene level to obtain a catalog of E3 ubiquitin ligase-encoded genes with gene identifiers as the basic unit includes: Obtain records of multiple candidate proteins belonging to the same gene identifier and their respective evidence status; Based on the evidence status record, a representative protein is selected from multiple candidate proteins belonging to the same gene identifier according to a preset priority. The preset priorities include: a first priority, selecting candidate proteins with a high-confidence boundary state label and excluding candidate proteins with an exclusion boundary state label; if multiple candidate proteins are in the same boundary state, a second priority, selecting the candidate protein with the highest core domain integrity index; if the integrity indices are the same, a third priority, selecting the candidate protein with the highest domain coverage; if the coverage is the same, a fourth priority, selecting the candidate protein with the most fused multi-sources; if they still cannot be distinguished, the audit records of the heterogeneous transcripts of multiple candidate proteins are retained, and the gene identifier is marked as requiring review. The domain coordinates, domain annotation evidence, core domain integrity indicators, and boundary state labels of the representative protein are used as the gene catalog records corresponding to its gene identifier.

[0007] Optionally, the step of fusing multiple source annotations determined to correspond to the same core structural domain, generating fused structural domain coordinates and recording the number of multiple sources fused, and determining the core structural domain integrity index based on the fused structural domain coordinates includes: For multiple source annotations that are determined to correspond to the same core structural domain, a set of merged structural domain start coordinates and structural domain end coordinates are generated by fusing the start coordinates and end coordinates of each source, and the number of multiple sources fused is recorded. The length of the core structural domain interval is calculated based on the starting and ending coordinates of the structural domain, and the ratio of this length to the preset target structural domain length is used to determine the structural domain coverage. The core domain integrity index is determined based on the presence of a core domain in the candidate protein, the domain coverage of the core domain, and the number of multi-source annotations fused to the core domain.

[0008] Optionally, based on the core domain integrity index, candidate proteins are determined, and evidence state records containing boundary state labels are generated, including: If the core domain of a candidate protein exists, the domain coverage reaches a preset threshold, and the exclusion domain exists in the state of "no exclusion domain exists", then it is marked as a high confidence state; the exclusion domain existence state is used to indicate whether the candidate protein contains a pre-specified non-target domain type. If a candidate protein does not meet the conditions for a high-confidence state and the number of fused multiple sources is lower than a preset threshold, it is marked as a cross-core boundary state. If the core domain of the candidate protein is absent, the domain coverage is lower than a preset threshold, or the exclusion domain exists, it is marked as an exclusion boundary state. The evidence state record is generated based on the boundary state labels of the candidate proteins.

[0009] Optionally, calculating the overlap and coordinate offset between different source structural domain intervals includes: Calculate the overlap length, overlap ratio, and coordinate offset between structural domain intervals from different sources; Wherein, the overlap length represents the length of the overlapping region between the two domain regions on the protein sequence coordinates; the overlap ratio represents the ratio of the overlap length to the shorter of the lengths of the domain regions being compared; and the coordinate offset represents the offset distance between the start positions and the offset distance between the end positions of the domain regions from different sources.

[0010] Optionally, the HMM domain search performed on the proteome sequence based on the domain model corresponding to the E3 ubiquitin ligase yields a candidate protein set including: The profile-HMM search program is invoked to perform domain searches on the entire proteome sequence and generate a domain hit result file. The protein identifier, domain start position, domain end position, domain expected value, score, and domain coverage in the search results are analyzed. Proteins whose domain expected values ​​meet a preset significance threshold and whose domain coverage meets a preset coverage threshold are included in the candidate protein set.

[0011] Optionally, obtaining the protein-transcript-gene mapping file includes: Obtain the genome annotation file of the target species; The header line of the proteome sequence file is parsed to extract the protein identifiers for each protein sequence; The gene field, transcript field, and protein-related field in the genome annotation file are parsed to extract gene identifiers, transcript identifiers, and their corresponding protein identifiers. The protein identifiers extracted from the genome annotation file and the protein identifiers extracted from the proteome sequence file are the same identifiers that exist in different files. Using the protein identifier as the association key, the gene identifier and transcript identifier extracted from the genome annotation file are mapped to the protein sequence in the proteome sequence file to generate the protein-transcript-gene mapping file.

[0012] Optionally, after performing HMM domain retrieval, the method further includes a similarity expansion retrieval step: Based on the reference E3 protein sequence, a similarity expansion search is performed on the proteome sequence to obtain a second set of candidate proteins; The first candidate protein set obtained by searching the HMM domains is merged with the second candidate protein set, and duplicate protein identifiers are removed to obtain the candidate protein set.

[0013] An E3 ubiquitin ligase-encoded gene recognition device, the device comprising: The acquisition module is used to acquire the proteome sequence file and protein-transcript-gene mapping file of the target species; The retrieval module is used to perform HMM domain retrieval on the proteome sequence based on the domain model corresponding to the E3 ubiquitin ligase to obtain a set of candidate proteins. The annotation module is used to obtain domain annotation results for each candidate protein in the candidate protein set from multiple domain database sources. The annotation results include domain type, start coordinates, end coordinates and source. The processing module is used to map the domain annotation results from multiple database sources to the protein sequence coordinate system of the same candidate protein, sort the domain intervals from different sources using the protein identifier as an index, calculate the overlap relationship and coordinate offset between the domain intervals from different sources, and determine whether the domain annotations from different sources correspond to the same core domain based on the overlap relationship and coordinate offset; for multiple annotations from different sources that are determined to correspond to the same core domain, they are fused to generate fused domain coordinates and record the number of multiple sources fused; based on the fused domain coordinates, the core domain integrity index is determined. The determination module is used to determine the candidate protein based on the core domain integrity index and generate an evidence state record containing boundary state labels. The determination module is used to merge multiple candidate proteins belonging to the same gene identifier into the gene level according to the protein-transcript-gene mapping file and evidence status record, so as to obtain a catalog of E3 ubiquitin ligase-encoded genes with gene identifier as the basic unit.

[0014] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the above-described method for recognizing E3 ubiquitin ligase-encoded genes.

[0015] The E3 ubiquitin ligase-encoded gene recognition method provided by this invention has the following beneficial effects: This invention aligns domain annotations from multiple databases from a scattered state to a unified protein sequence coordinate system by mapping them to the same coordinate system, calculating overlap relationships and coordinate offsets, and then objectively comparing domain intervals from different sources under a unified reference system. Based on overlap relationships and coordinate offsets, evidence pointing to the same domain from multiple sources is integrated into representative coordinates. Therefore, domain type determination and boundary definition no longer rely on the judgment of any single database, but are comprehensively evaluated based on the consistency of evidence from multiple sources under the same coordinate system, thus eliminating fluctuations in judgment caused by different database sources. Furthermore, the calculation of core domain integrity indicators and gene hierarchy merging are both completed under the aforementioned unified coordinate fusion benchmark, ensuring that the entire process from domain evidence integration to the final gene catalog output is based on unified domain coordinates, and the final output gene catalog remains stable and consistent under different database conditions. Attached Figure Description

[0016] To more clearly illustrate the embodiments and design schemes of the present invention, the accompanying drawings required for this embodiment will be briefly described below. The drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a flowchart illustrating an E3 ubiquitin ligase-encoded gene recognition method provided by the present invention according to an exemplary embodiment.

[0018] Figure 2 This is a flowchart illustrating another method for recognizing E3 ubiquitin ligase-encoded genes according to an exemplary embodiment of the present invention.

[0019] Figure 3 This is a block diagram of an E3 ubiquitin ligase-encoded gene recognition device provided by the present invention according to an exemplary embodiment. Detailed Implementation

[0020] To enable those skilled in the art to better understand and implement the technical solutions of the present invention, the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. The following embodiments are only used to more clearly illustrate the technical solutions of the present invention and should not be construed as limiting the scope of protection of the present invention.

[0021] The technical solutions provided by the various embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0022] First, this invention provides a method for recognizing E3 ubiquitin ligase-encoded genes, specifically as follows: Figure 1 As shown, it includes the following steps: S101. Obtain the proteome sequence file and protein-transcript-gene mapping file of the target species.

[0023] In this step, to obtain the protein-transcript-gene mapping file, it is necessary to obtain the genome annotation file of the target species; parse the header line of the proteome sequence file and extract the protein identifiers of each protein sequence; parse the gene field, transcript field, and protein-related field in the genome annotation file, and extract the gene identifier, transcript identifier, and their corresponding protein identifier. The protein identifiers extracted from the genome annotation file and the protein identifiers extracted from the proteome sequence file are the same identifiers existing in different files; using the protein identifiers as the association key, match the gene identifiers and transcript identifiers extracted from the genome annotation file with the protein sequences in the proteome sequence file to generate the protein-transcript-gene mapping file.

[0024] S102. Based on the domain model corresponding to E3 ubiquitin ligase, perform HMM domain search on the proteome sequence to obtain a set of candidate proteins.

[0025] In this step, the profile-HMM search program is invoked to perform domain search on the whole proteome sequence, generating a domain hit result file; the protein identifier, domain start position, domain end position, domain expected value, score, and domain coverage in the search results are parsed; proteins whose domain expected value meets the preset significance threshold and whose domain coverage meets the preset coverage threshold are included in the candidate protein set.

[0026] In addition, based on the reference E3 protein sequence, a similarity extension search can be performed on the proteome sequence to obtain a second candidate protein set; the first candidate protein set obtained by the HMM domain search is merged with the second candidate protein set, and duplicate protein identifiers are removed to obtain a candidate protein set.

[0027] S103. For each candidate protein in the candidate protein set, obtain the domain annotation results from multiple domain database sources.

[0028] The annotation results include the domain type, starting coordinates, ending coordinates, and source.

[0029] S104. Map the domain annotation results from multiple databases to the protein sequence coordinate system of the same candidate protein. Sort the domain intervals from different sources using the protein identifier as an index. Calculate the overlap and coordinate offset between the domain intervals from different sources. Determine whether the domain annotations from different sources correspond to the same core domain based on the overlap and coordinate offset. For multiple annotations from different sources that are determined to correspond to the same core domain, fuse them to generate fused domain coordinates and record the number of multiple sources fused. Determine the core domain integrity index based on the fused domain coordinates.

[0030] In this step, it is necessary to calculate the overlap length, overlap ratio, and coordinate offset between domain intervals from different sources. The overlap length represents the length of the overlapping region between two domain intervals on the protein sequence coordinates; the overlap ratio represents the ratio of the overlap length to the shorter of the lengths of the domain intervals being compared; and the coordinate offset represents the offset distance between the start and end positions of the domain intervals from different sources. For annotations from multiple sources that are determined to correspond to the same core domain, a set of fused domain start and end coordinates is generated based on the start and end coordinates of each source, and the number of fused sources is recorded. The domain interval length of the core domain is calculated based on the domain start and end coordinates, and the domain coverage is determined by the ratio to the preset target domain length. A core domain integrity index is determined based on whether the candidate protein has a core domain, the domain coverage of the core domain, and the number of annotations from multiple sources fused to that core domain.

[0031] In addition, for annotations with coordinate offsets exceeding the preset range or inconsistent domain types, the original coordinates are retained and conflict information is recorded as one of the verification conditions for candidate proteins in gene identifiers that need to be verified.

[0032] S105. Based on the core domain integrity index, the candidate proteins are judged and evidence state records containing boundary state labels are generated.

[0033] In this step, if the core domain of a candidate protein exists, the core domain integrity index reaches a preset threshold, the domain coverage reaches a preset threshold, and the exclusion domain presence status is "excluded domain not present," it is marked as a high-confidence state. The exclusion domain presence status is used to indicate whether the candidate protein contains a pre-specified non-target domain type. If the candidate protein does not meet the conditions for a high-confidence state, and the number of fused multi-sources is lower than a preset number threshold, it is marked as a verification boundary state. If the core domain of a candidate protein does not exist, the domain coverage is lower than a preset threshold, or the exclusion domain presence status is "excluded domain present," it is marked as an exclusion boundary state. An evidence status record is generated based on the boundary status labels of the candidate protein.

[0034] S106. Based on the protein-transcript-gene mapping file and evidence status record, multiple candidate proteins belonging to the same gene identifier are merged into the gene level to obtain a catalog of E3 ubiquitin ligase-encoded genes with gene identifiers as the basic unit.

[0035] In this step, multiple candidate proteins belonging to the same gene identifier and their respective evidence status records are obtained. Based on the evidence status records, a representative protein is selected from the multiple candidate proteins belonging to the same gene identifier according to a preset priority. The preset priority includes: first priority, selecting candidate proteins with a boundary status label of high confidence, and excluding candidate proteins with a boundary status label of exclusion boundary; if multiple candidate proteins are in the same boundary state, then second priority, selecting the candidate protein with the highest core domain integrity index; if the integrity index is the same, then third priority, selecting the candidate protein with the highest domain coverage; if the coverage is the same, then fourth priority, selecting the candidate protein with the most fused multi-sources; if they still cannot be distinguished, the audit records of heterogeneous transcripts of multiple candidate proteins are retained, and the gene identifier is marked as requiring review; the domain coordinates, domain annotation evidence, core domain integrity index, and boundary status label of the representative protein are used as the gene catalog record content corresponding to its gene identifier. Using the above method, domain annotations from multiple databases are mapped from a scattered state to the same protein sequence coordinate system. Overlap relationships and coordinate offsets are calculated, aligning domain regions from different sources to a unified reference system for objective comparison. Based on overlap relationships and coordinate offsets, evidence pointing to the same domain from multiple sources is integrated into representative coordinates, while retaining conflicting original coordinate records. Therefore, domain type determination and boundary delineation no longer rely on the judgment of any single database, but are comprehensively evaluated based on the consistency and conflict status of evidence from multiple sources within the same coordinate system, thus eliminating fluctuations in judgment caused by different database sources. Furthermore, the calculation of core domain integrity indicators and gene hierarchy merging are both completed under the same coordinate fusion benchmark, ensuring that the entire process from domain evidence integration to the final gene catalog output is based on unified domain coordinates, and the final output gene catalog remains stable and consistent across different databases.

[0036] Based on the above steps, the present invention also provides an embodiment for illustration.

[0037] This embodiment provides a method for identifying E3 ubiquitin ligase-encoded genes based on proteomic domain fusion analysis. For example... Figure 2 As shown, this scheme takes the target species' proteome sequence file and genome annotation file as input, and achieves computer-aided identification of E3 coding genes through domain model retrieval, similarity expansion, multi-database domain annotation parsing, domain coordinate fusion, boundary false positive resolution, multi-transcriptor compression, and gene-level directory output. Additionally, some keywords for subsequent steps are explained here for better understanding, as shown in Table 1.

[0038] Table 1 Keyword Definition Table S1 Data Acquisition and Preprocessing: Acquire the target species' proteome sequence file, optional CDS (coding sequence) file, GFF3 / GTF genome annotation file, protein-transcript-gene mapping file, species information table, and runtime parameter file. Perform unique ID checks, illegal character checks, length checks, and duplicate record checks on the protein sequences, and establish the mapping relationship between protein_id (protein identifier), transcript_id (transcript identifier), and gene_id (gene identifier).

[0039] The system reads the target species' protein FASTA file, optional CDS file, GFF3 / GTF annotation file, protein-transcript-gene mapping file, species information table, and runtime parameter file. The system parses the header line of the protein FASTA file, extracts the protein_id, and checks for duplicate protein_ids, empty sequences, illegal amino acid characters, abnormally short sequences, or abnormally long sequences. The system parses the gene, mRNA / transcript, CDS, or protein-related fields in the GFF3 / GTF file and establishes a mapping table between protein_id, transcript_id, and gene_id. For species whose mapping relationships cannot be directly obtained from GFF3 / GTF, the system reads an external mapping table as a supplement. Upon completion, the system outputs an input manifest file (an input file list or input file record table), a sequence statistics table, and an ID mapping table.

[0040] S2 Domain Model and Reference Sequence Reading: Reads the domain model file corresponding to the target E3 type and the set of reference E3 protein sequences. The parameter file includes core domain model numbers, auxiliary domains, excluded domains, significance thresholds, coverage thresholds, and similarity extension parameters. This parameter file is used to drive computer retrieval and is not protected as part of the human classification rules themselves.

[0041] The system reads the domain model file corresponding to the target E3 type and the set of reference E3 protein sequences. The domain model file includes core domain HMM profiles (Hidden Markov Model domain model files) for E3 types such as U-box / PUB, HECT, and RING / RBR. The runtime parameter file records the core domain model number, auxiliary domain name, excluded domain name, significance threshold, coverage threshold, similarity extension parameters, and output path. This parameter file is used to drive computer retrieval and result parsing and is not protected as part of the manually defined classification rules themselves.

[0042] S3 HMM domain search: Perform profile-HMM search on the whole proteome sequence to obtain the first set of candidate proteins, and record the domain hit interval, domain i-Evalue (domain independent E value or domain significance value), score, domain start / end (domain start and end positions) and domain coverage.

[0043] The system invokes the profile-HMM search program to perform domain searches on the entire proteome sequence, generating a domain hit result file. The system parses the protein_id, domain start position, domain end position, domain i-Evalue, score, bias score, HMM start and end positions, and alignment start and end positions from the search results, and calculates the domain coverage. Proteins meeting the significance and coverage requirements are added to the first candidate set; proteins that do not reach the threshold but are close to it are retained as weak evidence records for subsequent joint evaluation with similarity expansion and multi-database domain annotation results.

[0044] S4 Similarity Expansion Search: Based on the reference E3 protein sequence, a similarity expansion search is performed on the entire proteome sequence to obtain a second set of candidate proteins. Similarity expansion can be implemented by DIAMOND blastp or BLASTP, and the similarity_tool, e-value, identity, coverage, best_hit, and reference_family are recorded.

[0045] The system uses the reference E3 protein sequence as the query set to perform a similarity expansion search on the whole proteome sequences of the target species. This step can be implemented using DIAMOND blastp or BLASTP. The system parses the query_id (query sequence identifier), subject_id (target sequence identifier), identity (sequence consistency), alignmentlength, mismatch, gap open, query coverage / qcov (query sequence coverage), subject coverage (target sequence coverage), e-value (expected value or significance value), bit score, and best hit information in the output results, and adds the target protein corresponding to subject_id to the second candidate set. For large-scale cross-species proteome data, DIAMOND is preferred to improve operational efficiency.

[0046] S5 Candidate Set Merging: Merge the first and second candidate protein sets, remove duplicate protein_ids, and obtain the candidate protein set to be validated. This candidate set is calculated from the whole proteome, rather than being given manually in advance.

[0047] The system merges the first and second candidate sets by protein_id, removing duplicate protein_ids to form a set of candidate proteins to be validated. For proteins that are simultaneously detected by HMM and similarity expansion, the system retains both sources of evidence; for proteins detected only by a single path, the system retains their single source of evidence and marks them as candidates to be validated. This candidate set is calculated from the whole proteome and is not a list of candidate proteins pre-specified by humans.

[0048] S6 Multi-Database Domain Annotation Parsing: Performs InterProScan or equivalent domain annotation on candidate proteins to be validated, and parses domain types, start and end coordinates, significance values, coverage, and data sources from Pfam, SMART, PROSITE, InterPro, etc.

[0049] The system performs InterProScan or equivalent domain annotation on candidate proteins to be validated and parses domain annotation results from sources such as Pfam, SMART, PROSITE Profiles, and InterPro. The system extracts the domain type, domain accession (database accession number or entry number), domain start coordinates, end coordinates, significance value, annotation source, and database name, and stores them as a domain annotation table by protein_id. For cases where the same protein is annotated to the same or similar domains by multiple databases, the system retains all sources and proceeds to the domain coordinate fusion step.

[0050] S7 Domain Coordinate Fusion Analysis: This method maps domain annotation results from different sources to the same candidate protein coordinate system, and calculates domain overlap relationships, core domain integrity, auxiliary domain combination status, excluded domain presence status, and consistency of evidence from multiple domain sources.

[0051] The system maps domain annotation results from sources such as Pfam, SMART, PROSITE, and InterPro to the same protein coordinate system. The system sorts domain intervals by protein_id, calculates the overlap ratio, coverage, and coordinate offset between intervals from different sources, and determines whether they belong to the same target core domain. For domains supported by multiple sources, the system generates representative domain coordinates; for domains with significant coordinate differences between sources, the system retains the coordinates of each source and records conflicts. The system simultaneously calculates core domain integrity, auxiliary domain combination status, excluded domain presence status, and consistency of evidence from multiple domain sources.

[0052] For example, the system uses protein_id as an index to load domain annotation results from sources such as Pfam, SMART, PROSITE, and InterPro obtained by S6 into the same candidate protein coordinate system. The system reads the protein_id, domain_type, domain_source (domain annotation source), domain_start, domain_end, domain_evalue, and domain_coverage fields from each domain annotation record.

[0053] For multiple domain annotation intervals under the same protein_id, the system sorts the domain intervals according to domain_start and domain_end, and calculates overlap_length, overlap_ratio, and coordinate_offset between domain intervals from different sources. Here, overlap_length represents the overlap length of two domain intervals on the protein sequence coordinates; overlap_ratio represents the ratio between overlap_length and the shorter domain interval length or a preset reference length; and coordinate_offset represents the start and end position offset between the boundaries of annotation intervals from different sources.

[0054] When domain annotation intervals from different sources meet a preset overlap condition in coordinates, and their `domain_type` belongs to the same target core domain or equivalent domain category, the system merges them into a single evidence group for the same target domain and generates representative domain coordinates. For cases where domain coordinates from different sources differ significantly, domain types are inconsistent, or the system is supported by only a single source, the system retains the original source coordinates and records the `conflict_source` and `conflict_reason` fields.

[0055] This step generates the intermediate table domain_coordinate_fusion.tsv, which includes at least the following fields: protein_id, domain_type, domain_source, domain_start, domain_end, representative_domain_start, representative_domain_end, domain_coverage, overlap_ratio, multi_source_support_count (number of multiple source supports), core_domain_presence (presence status of core domains), core_domain_completeness (completeness of core domains), auxiliary_domain_presence (presence status of auxiliary domains), exclusion_domain_presence (presence status of excluded domains), conflict_source (source of conflict), and conflict_reason (reason of conflict).

[0056] S8 Boundary False Positive Resolution: Computer-based determination of candidate proteins based on insufficient domain coverage, missing core domains, strong similarity but lack of domain evidence, RING-like, NSMCE2-like, RBBP6-like, or other non-target domain combinations, and generation of evidence status records.

[0057] The system uses computer-based judgment to identify false positive candidates at the boundary based on the domain coordinate fusion results and similarity expansion results. The processed objects include boundary types such as insufficient domain coverage, missing core domains, strong similarity but lacking domain evidence, presence of non-target domain combinations, and RING-like, NSMCE2-like, and RBBP6-like boundaries. The system does not rely on human experience for this step; instead, it generates evidence status records for candidate proteins based on domain presence, domain coverage, domain combination status, and source consistency.

[0058] For example, the system reads the similarity search results generated in S4, the multi-database domain annotation results generated in S6, and the domain_coordinate_fusion.tsv generated in S7, and marks the boundary states of candidate proteins. This step is not based on manual experience, but rather on the combination of fields generated in the previous steps to generate candidate protein state records.

[0059] The system prioritizes processing the following types of candidates: those that are matched by similarity retrieval but lack a core structural domain; those that are matched by HMM retrieval but are not supported by multi-database structural domain annotations; those whose core structural domain coverage is below a preset threshold; those with exclusion_domain_presence; those with RING-like, NSMCE2-like, RBBP6-like, or other non-target structural domain combinations; and those with obvious structural domain coordinate conflicts that require verification.

[0060] The system generates `boundary_status` and `boundary_reason` based on the fields `core_domain_presence`, `core_domain_completeness`, `domain_coverage`, `multi_source_support_count`, `exclusion_domain_presence`, `similarity_evalue`, `query_coverage`, `subject_coverage`, `best_hit_family`, `conflict_source`, and `conflict_reason`. `boundary_status` can include states such as `high-confidence`, `review-boundary`, and `excluded-boundary`, representing the evidence status of candidate proteins. This status serves only as an intermediate state or final quality control label in the computer processing flow and is not protected by the human classification rules themselves.

[0061] This step generates the intermediate table boundary_candidate_status.tsv, which includes at least the protein_id, gene_id, boundary_type, supporting_fields, conflict_fields, boundary_status, and boundary_reason fields.

[0062] S9 Multi-transcript Compression and Representative Protein Selection: Based on the mapping relationship between protein_id, transcript_id, and gene_id, multiple transcripts or multiple protein results are compressed into gene_id-level results. Representative proteins are selected based on canonical transcript, primary transcript, domain integrity, protein length, or preset priority, while retaining isoform audit records.

[0063] The system groups multiple proteins or transcripts to the gene_id level based on the mapping relationship between protein_id, transcript_id, and gene_id. When the same gene_id corresponds to multiple protein_ids, the system selects a representative protein based on canonical transcript, primary transcript, core domain integrity, domain coverage, protein length, or a preset priority. The system also retains isoform audit records for other transcripts to avoid duplicate counting due to multiple transcripts.

[0064] For example, the system reads the protein_id-transcript_id-gene_id mapping table established in S1 and groups the candidate protein evidence generated in S3 to S8 according to gene_id. In cases where the same gene_id corresponds to multiple transcript_ids or multiple protein_ids, the system treats them as multiple transcripts or multiple protein records of the same gene, avoiding duplicate counting of multiple transcripts of the same gene in the final E3 coding gene catalog.

[0065] The system selects representative proteins according to preset priorities. These priorities include: canonical transcript priority, primary transcript priority, higher core domain completeness priority, higher domain coverage priority, higher multi-source support count priority, high-confidence boundary status priority, and more complete protein length priority. If different candidate proteins still cannot be distinguished under the above conditions, the system retains the isoform_audit_records of multiple candidate proteins and marks the gene_id as the gene-level record that needs to be reviewed.

[0066] This step generates the intermediate table gene_isoform_collapse.tsv, which includes at least the following fields: gene_id, representative_protein_id (representing protein identifier), representative_transcript_id (representing transcript identifier), all_mapped_protein_ids (all mapped protein identifiers), all_mapped_transcript_ids (all mapped transcript identifiers), isoform_count, representative_selection_reason (representing transcription / protein selection criteria), gene_level_boundary_status (gene level boundary status), and isoform_audit_record.

[0067] S10 Result Output and Audit Log: Outputs an E3 ubiquitin ligase encoding gene-level directory, including gene_id, transcript_id, protein_id, representative protein, domain type, domain coordinates, similarity search evidence, domain annotation evidence, quality control labels, running parameters, and log files. The quality control labels are only used as the final output status of the process to identify retention, review, or rejection; they are not part of the core invention of this application.

[0068] The system outputs a gene-level directory for the E3 ubiquitin ligase-encoding gene. The output includes gene_id, transcript_id, protein_id, representative protein ID, domain type, domain start and end coordinates, domain coverage, HMM search evidence, similarity search evidence, multi-database domain annotation evidence, quality control labels, runtime parameters, input manifest file, and log file. The quality control labels are only the final output status of the process, used to identify retention, review, or rejection, and are not part of the core invention of this application.

[0069] The data processing flow described above is shown in Table 2.

[0070] Table 2 Data Processing Flowchart Based on the input files, search results, intermediate tables, and runtime logs from S1 to S9, the system compiles and generates a gene-level directory of E3 ubiquitin ligase-encoding genes. The final output is not independent of the results generated in the preceding steps, but is obtained by summarizing, merging, and formatting the aforementioned intermediate tables according to gene_id. The data relationships are shown in Table 3.

[0071] Table 3 Data Relationship Table In this invention, the quality control tag serves only as the final output status of the process, identifying the evidence status of candidate genes or proteins, and is not part of the core invention. The domain retrieval parameter file, structured evidence record, and quality control tag all serve the computer-implemented data processing flow from whole proteome input to gene-level E3 coding gene catalog output. Furthermore, this invention provides the following quantitative fields and calculation indicators to illustrate the feasibility of this method, as shown in Table 4.

[0072] Table 4 Quantitative Fields and Calculation Indicators In some embodiments, HMMER 3.4 can be used for profile-HMM retrieval, DIAMOND 2.1.24 or BLASTP can be used for similarity expansion, and InterProScan 5.77-108.0 can be used to call Pfam, SMART and PROSITE Profiles for domain annotation.

[0073] Example 1: Identification of Astragalus U-box / PUB coding genes. Using the whole Astragalus proteome sequence, GFF3 annotation file, and protein-transcript-gene mapping as input, the U-box / PUB related domain model and reference protein sequence were read. HMM search and DIAMOND / BLASTP similarity expansion were performed. Pfam, SMART, PROSITE, and InterPro domain coordinates were analyzed. Evidence from multiple sources was fused and analyzed, and the domains were merged to the gene_id level according to the mapping file, forming a directory of 64 AmPUB gene-level genes for subsequent phylogenetic, structural, replication, expression trend, and qRT-PCR candidate gene priority analysis.

[0074] Example 2: Identification of multi-species U-box / PUB coding genes in animals. Using the Ensembl release 115 animal canonical transcript protein as input, DIAMOND was employed as the preferred implementation for large-scale similarity expansion. Combined with domain coordinate fusion, multi-transcript compression, and gene-level output, a multi-species U-box / PUB member directory was formed. This example illustrates that the present invention can be extended from single-species to multi-species, large-scale proteomic data.

[0075] Example 3: Expanded Recognition of HECT and RING / RBR Encoding Genes. For the HECT family, HECT domain models can be read and processed in conjunction with catalytic cysteine ​​sites, domain integrity, and length range. For the RING / RBR family, RING / RBR domain models can be read and processed in conjunction with domain combinations, zinc binding sites, and boundary domain states. Therefore, this invention is applicable not only to U-box / PUB but also to the recognition of encoding genes for major subclasses of E3 ubiquitin ligases.

[0076] Secondly, the present invention also provides an E3 ubiquitin ligase-encoded gene recognition device, such as... Figure 3 As shown, it includes: The acquisition module 201 is used to acquire the proteome sequence file and protein-transcript-gene mapping file of the target species.

[0077] The retrieval module 202 is used to perform HMM domain retrieval on the proteome sequence based on the domain model corresponding to the E3 ubiquitin ligase to obtain a set of candidate proteins.

[0078] The annotation module 203 is used to obtain domain annotation results for each candidate protein in the candidate protein set from multiple domain database sources. The annotation results include domain type, start coordinates, end coordinates and source.

[0079] The processing module 204 is used to map the domain annotation results from multiple database sources to the protein sequence coordinate system of the same candidate protein, sort the domain intervals from different sources using the protein identifier as an index, calculate the overlap relationship and coordinate offset between the domain intervals from different sources, and determine whether the domain annotations from different sources correspond to the same core domain based on the overlap relationship and coordinate offset; merge the annotations from multiple sources that are determined to correspond to the same core domain, generate the merged domain coordinates and record the number of multiple sources merged; and determine the core domain integrity index based on the merged domain coordinates.

[0080] The determination module 205 is used to determine the candidate protein based on the core structural domain integrity index and generate an evidence state record containing boundary state labels.

[0081] The determination module 206 is used to merge multiple candidate proteins belonging to the same gene identifier into the gene level according to the protein-transcript-gene mapping file, so as to obtain a catalog of E3 ubiquitin ligase-encoded genes with gene identifiers as the basic unit.

[0082] The present invention also provides a computer-readable storage medium storing a computer program that can be used to execute the above-described... Figure 1 The steps of a method for recognizing E3 ubiquitin ligase-encoded genes are provided.

[0083] This invention also provides a computer device. At the hardware level, the computer device includes a processor, an internal bus, a network interface, memory, and non-volatile memory, and may also include other hardware required for various operations. The processor reads the corresponding computer program from the non-volatile memory into memory and then executes it to achieve the above-mentioned functions. Figure 1 The steps of a method for recognizing E3 ubiquitin ligase-encoded genes are provided.

[0084] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0085] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, as well as combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0086] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0087] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0088] It should be noted that the specific embodiments described above enable those skilled in the art to more fully understand the present invention, but do not limit the present invention in any way. Therefore, although the present invention has been described in detail in this specification, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the present invention; and all technical solutions and improvements that do not depart from the spirit and scope of the present invention are covered within the protection scope of the patent of the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

Claims

1. A method for recognizing E3 ubiquitin ligase-encoded genes, characterized in that, The method includes: Obtain the proteome sequence file and protein-transcript-gene mapping file of the target species; Based on the domain model corresponding to E3 ubiquitin ligase, HMM domain search was performed on the proteome sequence to obtain a set of candidate proteins. For each candidate protein in the candidate protein set, domain annotation results are obtained from multiple domain database sources. The annotation results include domain type, start coordinates, end coordinates, and source. Domain annotation results from multiple databases are mapped to the protein sequence coordinate system of the same candidate protein. Domain intervals from different sources are sorted using the protein identifier as an index. The overlap and coordinate offset between domain intervals from different sources are calculated. Based on the overlap and coordinate offset, it is determined whether domain annotations from different sources correspond to the same core domain. For multiple annotations determined to correspond to the same core domain, they are fused to generate fused domain coordinates and the number of fused sources is recorded. Based on the fused domain coordinates, a core domain integrity index is determined. Based on the core domain integrity index, candidate proteins are determined, and evidence state records containing boundary state labels are generated. Based on the protein-transcript-gene mapping file and evidence status record, multiple candidate proteins belonging to the same gene identifier are merged into the gene level to obtain the E3 ubiquitin ligase-encoded gene catalog of the target species with gene identifier as the basic unit.

2. The method according to claim 1, characterized in that, The process of merging multiple candidate proteins belonging to the same gene identifier to the gene level to obtain a catalog of E3 ubiquitin ligase-encoded genes with gene identifiers as the basic unit includes: Obtain records of multiple candidate proteins belonging to the same gene identifier and their respective evidence status; Based on the evidence status record, a representative protein is selected from multiple candidate proteins belonging to the same gene identifier according to a preset priority. The preset priorities include: a first priority, selecting candidate proteins with a high-confidence boundary state label and excluding candidate proteins with an exclusion boundary state label; if multiple candidate proteins are in the same boundary state, a second priority, selecting the candidate protein with the highest core domain integrity index; if the integrity indices are the same, a third priority, selecting the candidate protein with the highest domain coverage; if the coverage is the same, a fourth priority, selecting the candidate protein with the most fused multi-sources; if they still cannot be distinguished, the audit records of the heterogeneous transcripts of multiple candidate proteins are retained, and the gene identifier is marked as requiring review. The domain coordinates, domain annotation evidence, core domain integrity indicators, and boundary state labels of the representative protein are used as the gene catalog records corresponding to its gene identifier.

3. The method according to claim 1, characterized in that, The process of fusing multiple source annotations that are determined to correspond to the same core structural domain, generating fused structural domain coordinates, and recording the number of multiple sources fused, and determining the core structural domain integrity index based on the fused structural domain coordinates, includes: For multiple source annotations that are determined to correspond to the same core structural domain, a set of merged structural domain start coordinates and structural domain end coordinates are generated by fusing the start coordinates and end coordinates of each source, and the number of multiple sources fused is recorded. The length of the core structural domain interval is calculated based on the starting and ending coordinates of the structural domain, and the ratio of this length to the preset target structural domain length is used to determine the structural domain coverage. The core domain integrity index is determined based on the presence of a core domain in the candidate protein, the domain coverage of the core domain, and the number of multi-source annotations fused to the core domain.

4. The method according to claim 3, characterized in that, Based on the core domain integrity index, candidate proteins are evaluated, and evidence state records containing boundary state labels are generated, including: If the core domain of a candidate protein exists, the domain coverage reaches a preset threshold, and the exclusion domain exists in the state of "no exclusion domain exists", then it is marked as a high confidence state; the exclusion domain existence state is used to indicate whether the candidate protein contains a pre-specified non-target domain type. If a candidate protein does not meet the conditions for a high-confidence state and the number of fused multiple sources is lower than a preset threshold, it is marked as a cross-core boundary state. If the core domain of the candidate protein is absent, the domain coverage is lower than a preset threshold, or the exclusion domain exists, it is marked as an exclusion boundary state. The evidence state record is generated based on the boundary state labels of the candidate proteins.

5. The method according to claim 1, characterized in that, The calculation of the overlap relationship and coordinate offset between structural domain intervals from different sources includes: Calculate the overlap length, overlap ratio, and coordinate offset between structural domain intervals from different sources; Wherein, the overlap length represents the length of the overlapping region between the two domain regions on the protein sequence coordinates; the overlap ratio represents the ratio of the overlap length to the shorter of the lengths of the domain regions being compared; and the coordinate offset represents the offset distance between the start positions and the offset distance between the end positions of the domain regions from different sources.

6. The method according to claim 1, characterized in that, The HMM domain search performed on the proteome sequence based on the domain model corresponding to E3 ubiquitin ligase yielded a candidate protein set including: The profile-HMM search program is invoked to perform domain searches on the entire proteome sequence and generate a domain hit result file. The protein identifier, domain start position, domain end position, domain expected value, score, and domain coverage in the search results are analyzed. Proteins whose domain expected values ​​meet a preset significance threshold and whose domain coverage meets a preset coverage threshold are included in the candidate protein set.

7. The method according to claim 1, characterized in that, Obtaining the protein-transcript-gene mapping file includes: Obtain the genome annotation file of the target species; The header line of the proteome sequence file is parsed to extract the protein identifiers for each protein sequence; The gene field, transcript field, and protein-related field in the genome annotation file are parsed to extract gene identifiers, transcript identifiers, and their corresponding protein identifiers. The protein identifiers extracted from the genome annotation file and the protein identifiers extracted from the proteome sequence file are the same identifiers that exist in different files. Using the protein identifier as the association key, the gene identifier and transcript identifier extracted from the genome annotation file are mapped to the protein sequence in the proteome sequence file to generate the protein-transcript-gene mapping file.

8. The method according to claim 1, characterized in that, After performing HMM structural domain retrieval, the method further includes a similarity expansion retrieval step: Based on the reference E3 protein sequence, a similarity expansion search is performed on the proteome sequence to obtain a second set of candidate proteins; The first candidate protein set obtained by searching the HMM domains is merged with the second candidate protein set, and duplicate protein identifiers are removed to obtain the candidate protein set.

9. A gene recognition device encoded by an E3 ubiquitin ligase, characterized in that, The device includes: The acquisition module is used to acquire the proteome sequence file and protein-transcript-gene mapping file of the target species; The retrieval module is used to perform HMM domain retrieval on the proteome sequence based on the domain model corresponding to the E3 ubiquitin ligase to obtain a set of candidate proteins. The annotation module is used to obtain domain annotation results for each candidate protein in the candidate protein set from multiple domain database sources. The annotation results include domain type, start coordinates, end coordinates and source. The processing module is used to map the domain annotation results from multiple database sources to the protein sequence coordinate system of the same candidate protein, sort the domain intervals from different sources using the protein identifier as an index, calculate the overlap relationship and coordinate offset between the domain intervals from different sources, and determine whether the domain annotations from different sources correspond to the same core domain based on the overlap relationship and coordinate offset; for multiple annotations from different sources that are determined to correspond to the same core domain, they are fused to generate fused domain coordinates and record the number of multiple sources fused; based on the fused domain coordinates, the core domain integrity index is determined. The determination module is used to determine the candidate protein based on the core domain integrity index and generate an evidence state record containing boundary state labels. The determination module is used to merge multiple candidate proteins belonging to the same gene identifier into the gene level according to the protein-transcript-gene mapping file and evidence status record, so as to obtain a catalog of E3 ubiquitin ligase-encoded genes with gene identifier as the basic unit.

10. A computer device, characterized in that, The method includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method described in any one of claims 1 to 8.

Images