A multi-genome consensus sequence high-speed identification method, system and application
Patent Information
- Application Number
- CN202510464238.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-04-14
- Publication Date
- 2026-10-20
AI Technical Summary
[0030]本发明的有益效果包括：本发明充分运用了基因组学数据以及相关方法、工具，能够在多基因组规模下高效的进行共有序列搜寻，能够针对物种全基因组核酸序列进行共有序列搜寻，或针对物种蛋白质序列进行共有序列搜寻，支持动态相似度阈值调整(30％-70％)，用于识别不同进化距离物种间的低相似度保守序列。
Smart Images
Figure BDA0005358033960000051 
Figure HDA0005358033970000011 
Figure HDA0005358033970000012
Abstract
Description
Technical Field
[0001] This invention belongs to the field of biotechnology, specifically relating to a high-speed identification method, system and application of multi-genome common sequences based on whole genome and protein sequence comparison, which is applicable to species evolution analysis, functional conserved element mining and cross-species gene function prediction. Background Technology
[0002] Shared sequence identification technology is a core tool for genome comparison and evolutionary analysis. Its goal is to identify conserved nucleic acid or protein sequences from the genomes or proteomes of different species. These shared sequences are of great value in revealing evolutionary relationships between species, identifying conserved functional elements (such as regulatory regions and structural domains), and developing cross-species molecular markers (such as universal primers).
[0003] However, existing shared sequence identification methods still face the following bottlenecks: Traditional methods rely on local alignment tools such as BLAST and MUMmer, which have high computational complexity (e.g., the O(n2) time complexity of BLAST), are too time-consuming to process large-scale genomic data, and are difficult to achieve rapid analysis of multiple species (e.g., more than 100 genomes), resulting in low efficiency and limited scale; Existing tools usually process nucleic acid or protein sequences independently, lacking the comprehensive analysis capability to integrate whole-genome nucleic acid fragments and protein domain information, making it difficult to identify complex conserved modules across sequence types (e.g., co-conserved regions of non-coding RNA and binding proteins); Most methods only support the statistical analysis of shared sequences at fixed taxonomic levels (e.g., at the species or genus level), and cannot dynamically adjust the taxonomic range according to research needs (e.g., screening by specific pathogen genus or host range), resulting in poor taxonomic flexibility and limiting their application in precision medicine and ecological research.
[0004] In recent years, improved methods based on k-mer indexes (such as Kraken) or deep learning (such as DeepHMM) have improved efficiency in some scenarios, but they still cannot effectively handle long fragments with low similarity. Furthermore, deep learning methods rely on a large amount of labeled data, have poor model generalization ability, and are difficult to adapt to new sequencing species.
[0005] To address the aforementioned issues, there is an urgent need for a high-speed method for identifying shared sequences across multiple genomes, capable of comparing whole genomes and protein sequences. This method should be able to perform high-speed parallel alignment, adapt to low-similarity sequences, and be compatible with multiple sequence types, thus meeting the need to find shared sequences in multiple genome sequences. Summary of the Invention
[0006] To address the shortcomings of existing technologies, the present invention aims to provide a high-speed method for identifying shared sequences across multiple genomes. This method can rapidly search for shared sequences targeting partial or complete nucleic acid sequences of a species, as well as one or more protein sequences within the species. The shared sequences screened by this method across different genomes can meet varying similarity requirements and can uncover shared sequences that are common to each other but have high mutation rates.
[0007] The proposed method for high-speed identification of shared sequences across multiple genomes includes: constructing a database of genomes and protein sequences from multiple species; constructing a pre-aligned sequence library for the target species; sequence alignment; processing of sequence alignment results; and detection and output of shared sequences.
[0008] The method for identifying shared sequences across multiple genomes according to the present invention specifically includes the following steps:
[0009] Step 1: Construct a multi-species genome and protein sequence target database: After clustering the genome and protein sequences of common species currently in the target database, construct a non-redundant target database G_1.
[0010] Step 2: Construct a pre-aligned sequence library for the target species: Divide the whole genome sequence of the target species under study into overlapping segments according to a certain length, and then cluster the divided nucleic acid and protein sequences to construct a pre-aligned sequence library G_2.
[0011] Step 3: Sequence alignment: Using sequence alignment software, the constructed pre-aligned sequence library G_2 and the non-redundant target database G_1 are aligned, and the alignment results are recorded.
[0012] Step 4: Sequence alignment result processing: Set constraints on the sequence alignment results according to specific criteria, including sequence similarity, sequence coverage, and sequence length, and filter out the matching sequences.
[0013] Step 5: Detection and output of common sequences: The selected alignment sequences are classified and grouped according to species classification requirements. Common sequences are then selected and output according to different species classification levels, which are the sequences common to the genomes of the expected species groups.
[0014] In a specific implementation plan, when a particular species is used as the detection target, the target sequence to be compared can be the entire genome of the species, or a specific genomic region or functional protein sequence of that species.
[0015] In step one, the genomic sequence data and protein sequence data used for clustering are sourced from the NCBI public database;
[0016] Specifically, the non-redundant target database G_1 refers to the non-redundant target database formed by integrating the post-translational amino acid sequences from the NCBI public database, including nucleic acid sequence databases such as core_nt and nt, after translation through six reading frames, and protein sequences from protein databases such as nr_clustered and nr, and then performing clustering using the CD-HIT tool (only the longest representative sequence is retained for sequences with similarity ≥90%) to remove redundancy;
[0017] By using six reading frames for translation, codon identification and amino acid conversion can be performed on the double strand from three starting points to fully cover the potential coding regions.
[0018] In step two, the method for constructing the alignment sequence library G_2 involves dividing the target genome nucleic acid sequence into sliding windows of 500-2000 bp in length (step size is 10% of the window length), using the k-mer clustering algorithm (k=10) to cluster the segmented sequences, and merging similar fragments (similarity ≥85%); or, dividing the target protein sequence into domains using databases including Pfam, and finally integrating the segmented nucleic acid sequences and protein sequences to form the pre-aligned sequence library G_2.
[0019] In step three, the sequence alignment mainly uses the DIAMOND algorithm. For nucleic acid sequences in the pre-aligned sequence library G_2, the DIAMOND blastx algorithm is used to align them to the non-redundant target database G_1; for protein sequences in the pre-aligned sequence library G_2, the DIAMOND blastp algorithm is used to align them to the non-redundant target database G_1, with e-value set to ≤ 1e-5, and multi-threading is started to accelerate the process.
[0020] In step four, the sequence alignment result processing refers to filtering candidate sequences based on preset sequence similarity thresholds (≥30%), sequence coverage (≥80%), and minimum sequence length (≥150bp or ≥50aa), and retaining sequences that meet the filtering conditions.
[0021] In step five, the species classification and tracing method for shared sequences refers to labeling the species origin of matching sequences based on the NCBI Taxonomy database, statistically analyzing the distribution frequency of shared sequences according to preset classification levels (such as genus, order, etc.), and selecting target shared sequences according to preset species groups (such as ≥10 shared species in mammals); outputting the shared sequences that meet the conditions, including outputting the species information where the shared sequences exist, and / or the location, length, similarity information, etc. of the shared sequences in the genome of each species.
[0022] The present invention also provides an identification system for implementing the above identification method, the identification system comprising: a database construction module, a pre-alignment sequence construction module, an alignment execution module, a result filtering module, a species tracing and classification module, and a result output module;
[0023] The database construction module is used to obtain the genomic nucleic acid sequences and protein sequences of multiple species from public databases, perform six reading frame translations on the nucleic acid sequences, and merge them with the protein sequences. Then, a clustering algorithm is used to remove redundancy and construct a non-redundant target database.
[0024] The pre-aligned sequence construction module is used to divide the genomic nucleic acid sequence of the target species into sliding windows, merge similar fragments using the k-mer clustering algorithm, and segment the protein sequence according to structural domains to construct a pre-aligned sequence library for the target species.
[0025] The alignment execution module is used to perform high-performance alignment of the pre-aligned sequence library and the non-redundant target database using the DIAMOND algorithm; wherein nucleic acid sequences are in blastx mode and protein sequences are in blastp mode, and multi-threaded parallel computing and e-value threshold setting are supported.
[0026] The result filtering module is used to compare and filter the results according to preset sequence similarity, sequence coverage and sequence length thresholds, and retain candidate common sequences;
[0027] The species origination and classification module is used to annotate candidate sequences based on the NCBI Taxonomy database and to filter common sequences according to the set classification level and species group conditions.
[0028] The result output module is used to output the common sequences that meet the conditions, as well as information including the species in which the common sequence is located, its position in the genome, sequence length, and alignment similarity.
[0029] This invention also provides applications of the above-mentioned identification method or identification system in species phylogenetic tree construction, functional annotation enhancement, conserved regulatory region prediction, and cross-species inference analysis of functional genes.
[0030] The beneficial effects of this invention include: This invention makes full use of genomic data and related methods and tools, enabling efficient search for common sequences at the multi-genome scale, and enabling search for common sequences for the whole genome nucleic acid sequence of a species or for the protein sequence of a species. It supports dynamic similarity threshold adjustment (30%-70%) and is used to identify low-similarity conserved sequences between species at different evolutionary distances. Attached Figure Description
[0031] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, 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.
[0032] Figure 1 This is a flowchart of a novel common sequence identification method according to the present invention.
[0033] Figure 2 This is a phylogenetic tree and a schematic diagram of the sequence transfer process of the shared sequences between bats and viruses discovered in the embodiments of the present invention.
[0034] Figure 3 This is a distribution diagram of the number of shared protein sequences between parasites, hosts, and symbiotic microorganisms in parasite species, as discovered in the embodiments of this invention. Detailed Implementation
[0035] The present invention will be further described in detail below with reference to the specific embodiments and accompanying drawings. Except for the contents specifically mentioned below, the processes, conditions, and experimental methods for implementing the present invention are all common knowledge and general knowledge in the art, and the present invention does not have any particular limitations.
[0036] This invention provides a novel high-speed method for identifying shared sequences across multiple genomes. This method can identify shared sequences between multiple genomes and other species, targeting both proteomic sequences and whole-genome nucleic acid sequences. The patent includes: constructing a target database of multi-genome nucleic acid and protein sequences; dividing and clustering the target genome and protein sequences to construct pre-aligned sequences; aligning the pre-aligned sequences to the target database using sequence alignment methods; performing similarity screening, splicing, species origin identification, and classification on the aligned sequences; and outputting results of shared nucleic acid or protein sequences across different species according to different requirements. Because this method can rapidly identify sequences in all currently sequenced species and can identify shared sequences with low similarity between different organisms (including whole-genome nucleic acid and protein sequences), it greatly improves the breadth and possibility of shared sequence identification and reduces the possibility of false negatives due to the inability to identify low-similarity sequences caused by sequence variations between different species.
[0037] Example 1: Search for shared sequences between bats and non-mammals
[0038] The steps for searching for shared sequences between bat genomes and other species are as follows:
[0039] (1) Constructing a multi-species genome and protein sequence database:
[0040] In this embodiment, the NT database downloaded from NCBI's FTP on July 18, 2020, and genomic data of 261,003 species downloaded from NCBI's reference genome database were used, and then clustered according to 90% similarity. The clustered nucleic acid sequences were translated according to six reading frames to obtain the translated amino acid sequences, and a non-redundant reference alignment library G_1 was constructed using these translated sequences.
[0041] (2) Construct a pre-aligned sequence library for the target species:
[0042] In this embodiment, six bat genomes with high sequencing quality were selected as target species (see Table 1). For the nucleic acid sequences of the six genomes, the genomes were segmented using a 1000bp sliding window with a step size of 200bp, and the segmented fragment set was used as the bat pre-aligned sequence library G_2.
[0043] Table 1. Information related to the genomes of six bat species.
[0044]
[0045] (3) Sequence alignment:
[0046] In this embodiment, the Diamond BlastX algorithm is used to align the bat sequence library G_2 to the reference alignment library G_1 (a non-redundant target database). The parameters "--evalue 1e-10-k 500--threads 32--outfmt 6" are used, which means that each sequence in the bat database is required to be strictly compared pairwise with the sequences in the reference database. 32 threads are used for acceleration. The output results will have a maximum of 500 alignment results for each sequence. The output format is a tab-separated blast result format (this format records the position information, similarity information, and length information of the two sequences).
[0047] (4) Sequence alignment result processing:
[0048] To broadly identify common sequences between bats and non-mammals, we selected those with a similarity greater than 30% and a length greater than 150 bp. We then spliced together overlapping sequences and clustered them to obtain the preliminary sequence alignment results.
[0049] (5) Detection and output of common sequences:
[0050] This embodiment uses species-related information from the NCBI taxonomy database to annotate the species information of each sequence in the preliminary sequence alignment results. Since the desired sequence is a shared sequence between bats and other non-mammals, based on the annotated sequence species information, for each bat sequence, the proportion of mammals (excluding Chiroptera) in the first 500 non-bat sequences aligned to it must be less than 5%. In this way, shared sequences existing between bats and non-mammals but not widely present in mammals can be found.
[0051] Regarding the results, this embodiment found 318 non-redundant shared sequences that met the criteria in six bat genomes, mainly found in different organisms such as bacteria, viruses, fish, and insects. Figure 2 As shown, a shared sequence between bats and viruses is presented. Phylogenetic tree (A) reveals that this sequence was likely transferred from the virus to the bat genome via horizontal gene transfer (B). This sequence, the HGT sequence of bat adeno-associated virus (HLrhiFer5:scaffold_m29_p_26:5204941~5205372), has been identified as a novel predicted gene coding region in the *Rhinolophus chinensis* genome, which is significant for understanding viral infection mechanisms and biological evolution.
[0052] Example 2: Search for uniquely shared functional protein sequences between parasites, their hosts, and symbiotic microorganisms
[0053] (1) Constructing a multi-species genome and protein sequence database:
[0054] In this embodiment, the NR database downloaded from NCBI's FTP on July 18, 2020, was used to construct a non-redundant protein sequence reference database G_1 after clustering based on 90% similarity.
[0055] (2) Construct a pre-aligned sequence library for the target species:
[0056] In this embodiment, 105 complete protein sequences of parasites were downloaded from the Protein database of NCBI on October 5, 2023, totaling 3,110,363 protein sequences. After clustering according to 90% similarity, 148,975 non-redundant parasite protein sequences were obtained. These sequences were used to construct the sequence library G_2 to be compared.
[0057] (3) Sequence alignment:
[0058] In this embodiment, the Diamond BLAST algorithm is used to align the parasite functional protein sequence library G_2 to the reference alignment library G_1. The parameters "--evalue 1e-5-k 1000--threads 16--outfmt 6" are used, which means that each protein sequence in the parasite is required to be paired with the sequence in the reference database. 16 threads are used for acceleration. The output results will output at most 1000 alignment results for each sequence. The output format is tab-separated BLAST result format (this format records the position information, similarity information and length information of the two sequences).
[0059] (4) Sequence alignment result processing:
[0060] In order to find common sequences between parasites and non-mammals more broadly, the results with similarity greater than 30% and length greater than 50aa were selected from the alignment results. Finally, the results were clustered to obtain the preliminary sequence alignment results.
[0061] (5) Detection and output of common sequences:
[0062] In this embodiment, the preliminary sequence alignment results are annotated with species-related information from the NCBI taxonomy database, providing species information for each sequence. Since this embodiment requires sequences of shared functional proteins between parasites, their hosts, and symbiotic microorganisms, based on the annotated sequence species information, for each parasite protein sequence, the number of non-parasite sequences aligned to the first 500 sequences must exceed 90%, excluding the corresponding host sequences and sequences from microorganisms. This allows for the identification of unique shared functional protein sequences between parasites, their hosts, and symbiotic microorganisms.
[0063] Regarding the results, a total of 3192 unique shared protein sequence families were found in the protein sequences of 105 parasitic species, and their distribution is as follows: Figure 3 As shown, nearly 70% (2119 sequences) of these sequences are shared by both species, particularly bacteria. Most of these sequences are functional proteins closely related to the parasite's catalytic processes and cellular metabolic regulation, suggesting that the symbiotic microenvironment may play a significant role in the parasite's life processes. In addition, 81 sequences were found to be shared by both the parasite and its corresponding host, and these sequences may have acquired their biological significance during the parasitic process and could contribute to parasitic specificity.
[0064] The scope of protection of this invention is not limited to the above embodiments. Any variations and advantages that can be conceived by those skilled in the art without departing from the spirit and scope of this invention are included in this invention and are protected by the appended claims.
Claims
1. A method for high-speed identification of shared sequences across multiple genomes, characterized in that, The identification method is based on whole-genome and protein sequence comparison and includes the following steps: Step 1: After clustering the genome sequences and protein sequences of the species in the target database, construct a non-redundant target database G_1; Step 2: Divide the whole genome sequence of the target species into overlapping segments, and then cluster the divided nucleic acid and protein sequences to construct a pre-aligned sequence library G_2; Step 3: Use sequence alignment software to align the constructed pre-aligned sequence library G_2 and the non-redundant target database G_1, and record the alignment results. Step 4: Set constraints on the sequence alignment results, including sequence similarity, sequence coverage, and sequence length, and filter out the matching sequences; Step 5: Perform species classification and grouping on the screened alignment sequences, and output common sequences according to different species classification level requirements.
2. The identification method as described in claim 1, characterized in that, In step one, the non-redundant target database G_1 is a non-redundant target database formed by integrating the post-translational amino acid sequences from the core_nt and nt nucleic acid sequence databases in the NCBI public database after translation through six reading frames, and the protein sequences from the nr_clustered and nr protein databases, and then clustering them using the CD-HIT tool to remove redundancy. During clustering, when the similarity of sequences is ≥90%, only the longest representative sequence is retained.
3. The identification method as described in claim 1, characterized in that, In step two, the method for constructing the alignment sequence library G_2 involves dividing the target genome nucleic acid sequence into sliding windows of 500-2000 bp in length, using the k-mer clustering algorithm to cluster the segmented sequences, and merging similar fragments. Alternatively, the target protein sequence can be segmented into domains using the Pfam database, and the segmented nucleic acid and protein sequences can be integrated to form a pre-aligned sequence library G_2. And / or, The sliding step size is 10% of the window length; in the clustering algorithm, the k value is 10; the similarity of similar segments used for merging is ≥85%; And / or, The sequences to be compared for the target species include the species’ whole genome, genomic regions, or functional protein sequences.
4. The identification method as described in claim 1, characterized in that, In step three, sequence alignment is performed using the DIAMOND algorithm. For nucleic acid sequences in the pre-aligned sequence library G_2, the DIAMOND blastx algorithm is used to align them to the non-redundant target database G_1; for protein sequences in the pre-aligned sequence library G_2, the DIAMOND blastp algorithm is used to align them to the non-redundant target database G_1. The e-value is set to ≤1e-5, and multi-threading is started to accelerate the alignment.
5. The identification method as described in claim 1, characterized in that, In step four, the sequence alignment results are processed based on preset sequence similarity thresholds, sequence coverage, and minimum sequence length to filter candidate sequences; And / or, The sequence similarity threshold is ≥30%, the sequence coverage is ≥80%, and the minimum sequence length is ≥150bp or ≥50aa.
6. The identification method as described in claim 1, characterized in that, In step five, the species classification and origin tracing method for shared sequences is based on the NCBI Taxonomy database to label the species origin of the matching sequences, count the distribution frequency of shared sequences according to preset classification levels, and screen out target shared sequences according to preset species groups; the shared sequences that meet the conditions are output, including the species information where the shared sequences exist, and / or the location, length, and similarity information of the shared sequences in the genome of each species.
7. The identification method as described in claim 1, characterized in that, The identification method searches for common sequences in the whole genome nucleic acid sequence of a species, or searches for common sequences in the protein sequence of a species.
8. The identification method as described in claim 1, characterized in that, The identification method supports dynamic similarity threshold adjustment, which is used to identify low-similarity conserved sequences between species at different evolutionary distances; And / or, The similarity threshold includes 30%-70%.
9. An identification system implementing the identification method as described in any one of claims 1-8, characterized in that, The identification system includes: a database construction module, a pre-alignment sequence construction module, an alignment execution module, a result filtering module, a species tracing and classification module, and a result output module; The database construction module is used to obtain the genomic nucleic acid sequences and protein sequences of multiple species from public databases, perform six reading frame translations on the nucleic acid sequences, and merge them with the protein sequences. Then, a clustering algorithm is used to remove redundancy and construct a non-redundant target database. The pre-aligned sequence construction module is used to divide the genomic nucleic acid sequence of the target species into sliding windows, merge similar fragments using the k-mer clustering algorithm, and segment the protein sequence according to structural domains to construct a pre-aligned sequence library for the target species. The alignment execution module is used to perform high-performance alignment of the pre-aligned sequence library and the non-redundant target database using the DIAMOND algorithm; wherein nucleic acid sequences are in blastx mode and protein sequences are in blastp mode, and multi-threaded parallel computing and e-value threshold setting are supported. The result filtering module is used to compare and filter the results according to preset sequence similarity, sequence coverage and sequence length thresholds, and retain candidate common sequences; The species origination and classification module is used to annotate candidate sequences based on the NCBI Taxonomy database and to filter common sequences according to the set classification level and species group conditions. The result output module is used to output the common sequences that meet the conditions, as well as information including the species in which the common sequence is located, its position in the genome, its sequence length, and its alignment similarity.
10. The application of the identification method as described in any one of claims 1-8, or the identification system as described in claim 9, in species phylogenetic tree construction, functional annotation enhancement, conserved regulatory region prediction, and cross-species inference analysis of functional genes.