Application method of sequence search tool CircBLAST considering gene sequence evolution rearrangement

Abstract

The application discloses an application method of a sequence search tool CircBLAST considering gene sequence evolution rearrangement, and belongs to the technical field of bioinformatics. The method flow comprises the following steps: firstly, all protein sequences are cut according to the length of a required word_size, a data set is constructed in combination with sequence annotation data, and is written into a database; then, a request sequence is prepared, and is cut into small fragments of word_size; further, a search matching, construction of a circular sequence and sequence alignment are carried out to complete a retrieval process; finally, an alignment result containing matching fragments, a similarity score and the like information is generated, is used for presenting to a user for viewing and judging the reliability of matching. The application considers the evolution rearrangement of gene sequences, significantly improves the accuracy of sequence alignment, and can find more reordered sequences.

Description

TECHNICAL FIELD

[0001] The application relates to an application method of a sequence search tool CircBLAST considering evolutionary rearrangement of gene sequences, and belongs to the technical field of bioinformatics. BACKGROUND

[0002] With the advancement of sequencing technology, the scale of protein / gene data in databases is becoming larger and larger. It is particularly important to analyze and annotate such a large data resource. In gene or protein data analysis, BLAST is a frequently used sequence alignment analysis tool. However, there are some problems in the existing BLAST method that need to be solved:

[0003] (1) Unable to effectively handle the rearrangement of gene or protein sequences: The traditional BLAST method is only suitable for forward sequence alignment and cannot correctly identify and compare sequence rearrangement. In the evolutionary process, gene duplication may lead to rearrangement of gene sequences, resulting in changes in protein sequences. Such sequence rearrangement is difficult to be discovered or compared in traditional BLAST.

[0004] (2) Lack of systematic study and database support for cyclic permutation: The existence and characteristics of cyclic permutation are a field that researchers have only recently begun to focus on. Although some researchers have systematically studied sequence rearrangement and established a cyclic permutation database (CPDB), there is a lack of specialized methods and tools for cyclic permutation.

[0005] Therefore, the current BLAST method has limitations in handling the rearrangement of gene or protein sequences, and lacks specialized methods and tools to accurately identify and compare cyclic permutations. These problems need to be solved to improve the accuracy and comprehensiveness of gene or protein data analysis. SUMMARY

[0006] The application provides an application method of a sequence search tool CircBLAST considering evolutionary rearrangement of gene sequences to solve the problems that the current BLAST method has limitations in handling the rearrangement of gene or protein sequences, and lacks specialized methods and tools to accurately identify and compare cyclic permutations.

[0007] The application provides an application method of a sequence search tool CircBLAST considering evolutionary rearrangement of gene sequences, and the application method of the sequence search tool CircBLAST considering evolutionary rearrangement of gene sequences comprises the following steps:

[0008] S100, database construction: according to the required word_size length, all protein sequences are cut, and a k-mers data set is constructed in combination with sequence annotation information; a key-value type data table is constructed according to the association relationship, and the key-value type data table is written into a database;

[0009] S200, Request sequence preparation: converting the protein sequence file to be compared into a required format, and then cutting the sequence into small pieces of length word_size, i.e. seed;

[0010] S300, Alignment process: comparing the request sequence with the sequences in the database, CircBLAST uses the SmithWaterman local alignment algorithm to search for partial matches of the request sequence in the database, and completes three steps of search matching, constructing a circular sequence and sequence alignment, and calculates the similarity and statistical significance of the match;

[0011] S400, Generate alignment results: CircBLAST returns the alignment results in the form of a list, which includes the matched sequence fragments and similarity scores, allowing users to evaluate the reliability of the match according to the similarity scores and statistical significance.

[0012] Further, in S100, the database construction includes the following steps:

[0013] S110, Cutting protein sequences into seed blocks: reading protein sequence data from the protein database, cutting each protein sequence into seed blocks of a fixed length according to the specified word_size, and dividing each seed block into a length of word_size in an overlapping or non-overlapping manner;

[0014] S120, Constructing a k-mers dataset: extracting key information from protein sequence annotation data; associating each seed block with the corresponding annotation information for subsequent database construction and query operations; based on the associated seed blocks and annotation information, constructing a k-mers dataset;

[0015] S130, Constructing a key-value data table: designing the structure of the data table, determining the keys and values in the table according to the seed blocks and associated annotation information, creating a key-value data table for storing seed blocks and annotation information, writing the associated seed blocks and annotation information into the data table, and writing the data table into the database.

[0016] Further, in S200, the following steps are included:

[0017] S210, Protein sequence file preparation: obtaining the protein sequence file to be compared;

[0018] S220, Sequence file format conversion: converting the protein sequence file into a required format;

[0019] S230, Seed block generation: cutting the converted protein sequence into small pieces of length word_size, i.e. seed blocks.

[0020] Furthermore, S230 includes the following steps:

[0021] S231. Traverse the sequence: For each protein sequence, traverse it starting from the beginning of the sequence according to the specified word_size;

[0022] S232, Seed block cutting: Starting from the current position, extract a segment of length word_size to form a seed block;

[0023] S233. Store seed block: Store the generated seed block in a data structure for subsequent cyclic sequence construction and comparison operations.

[0024] Furthermore, in S300, the following steps are included:

[0025] S310. Search and Match: CircBLAST calculates the index information of the hidden k-mers in each sequence using a mask. Using this index information, the requested sequence is compared with the database to find possible matching sequence fragments. The set of matching k-mers is called a Hit. If there are multiple consecutive Hits, they are called Hits.

[0026] S320. Construct circular sequences: Record the length of Hits for each sequence and the request sequence. Select the longest subsequence from all Hits as the starting point for the request sequence and the alignment sequence. Extend the selected starting sequence and the alignment sequence until the last amino acid residue before the subsequence is reached.

[0027] S330. Sequence alignment: The Smith-Waterman algorithm is used to perform sequence alignment. The alignment process considers sequence similarity, match length, and statistical significance of the match.

[0028] Furthermore, in S310, the index information of k-mers in each sequence is hidden by calculating the mask.

[0029] Furthermore, in S320, the length of the Hits of each sequence and the request sequence is recorded, and the subsequence with the longest Hits is selected as the starting point of the request sequence and the alignment sequence, extending the two sequences to the last amino acid residue before the subsequence.

[0030] Furthermore, in S330, the alignment process includes the alignment of the original sequence and the circular sequence, and simultaneously includes matching fragments and similarity score information in both cases.

[0031] Furthermore, the S400 includes the following steps:

[0032] S410. Alignment result data structure: CircBLAST returns alignment results in list form, with each alignment result item including the matched sequence fragment and similarity score;

[0033] S420. Extract sequence fragments: Extract matching sequence fragments from the circular sequence obtained during the alignment process. The sequence fragments are the fragments in the matching sequence that correspond to the request sequence, describing the similarity and matching relationship between the two sequences.

[0034] S430. Calculate similarity score: Calculate a similarity score for each alignment result to measure the degree of similarity between the aligned segments;

[0035] S440. Add statistically significant information: Add statistically significant information to the comparison results. Statistically significant information includes the expected value (E-value) and confidence level of the comparison.

[0036] S450, Organizational Comparison Results: Users judge the reliability of the match based on similarity scores and statistical significance.

[0037] Furthermore, the application method of CircBLAST, a sequence search tool that takes into account gene sequence evolution rearrangement, was used for gene sequence search.

[0038] The beneficial effects of this invention are as follows: Traditional BLAST search tools often fail to consider evolutionary rearrangements of gene sequences during the matching process. CircBLAST addresses this issue by effectively handling gene sequence evolution and rearrangements. By segmenting protein sequences and associating them with annotation information, CircBLAST constructs an efficient k-mers dataset and key-value data tables, improving database construction efficiency and query speed. CircBLAST employs the Smith-Waterman local alignment algorithm to find subsequences that partially match the requested sequence and calculates the similarity and statistical significance of the matches, providing more accurate and reliable alignment results. The alignment results include similarity scores and statistical significance information, allowing users to evaluate the reliability of the matches and further improve the credibility of the alignment results. Attached Figure Description

[0039] Figure 1 This is a flowchart of a method for applying CircBLAST, a sequence search tool that takes into account gene sequence evolution rearrangement, according to the present invention.

[0040] Figure 2 This is a schematic diagram of the construction of the circular sequence of the present invention;

[0041] Figure 3This is a schematic diagram of the data processing flow of the present invention, taking AAU08014.2 as an example;

[0042] Figure 4 This invention compares the domain distribution of BLAST and CircBLAST. In BLAST processing, S1 is the domain distribution of the AAU08014.2 sequence, and S2 is the domain distribution of the AMV61683.1 sequence. In CircBLAST processing of the cyclic sequence at the loop point, S3 is the domain distribution of the AAU08014.2 sequence, and S4 is the domain distribution of the AMV61683.1 sequence.

[0043] Figure 5 This is a box plot comparing the results of CircBLAST from this invention with those of four other software programs (Diamond, blast, mmseqs, hmmer). Detailed Implementation

[0044] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0045] In the accompanying drawings of specific embodiments of the present invention, in order to better and more clearly describe the working principle of each component in the system and show the connection relationship of each part in the device, only the relative positional relationship between each component is clearly distinguished. It does not constitute a limitation on the signal transmission direction, connection sequence, or size, dimension, and shape of each part within the component or structure.

[0046] A method for applying CircBLAST, a sequence search tool that also considers gene sequence evolution rearrangement, includes the following steps:

[0047] S100. Database Construction: Based on the required word_size length, all protein sequences are segmented, and a k-mers dataset is constructed by combining sequence annotation information. A key-value data table is constructed based on the association relationship, and the key-value data table is written into the database.

[0048] S200, Request Sequence Preparation: Convert the protein sequence file to be aligned into a format that meets the requirements, and then cut the sequence into small fragments of length word_size, i.e., seeds;

[0049] S300, Alignment Process: The requested sequence is aligned with the sequences in the database. CircBLAST uses the Smith-Waterman local alignment algorithm to search for subsequences in the database that partially match the requested sequence. It completes three steps: search and matching, construction of circular sequences, and sequence alignment, and calculates the similarity and statistical significance of the matches.

[0050] S400. Generate alignment results: CircBLAST returns the alignment results in list form, including matched sequence fragments and similarity scores, allowing users to assess the reliability of the match based on the similarity scores and statistical significance.

[0051] Specifically, S100, Database Construction: Based on the required word size, all protein sequences are segmented, and a k-mers dataset is constructed by combining sequence annotation information. A key-value table is then built based on the association relationships and written into the database. This database construction method offers efficient indexing and querying speeds, more accurate association between protein sequences and annotation information, and a modular design that makes database construction and updates more flexible and scalable.

[0052] S200, Sequence Preparation Request: Convert the protein sequence file to be aligned into a compliant format, and simultaneously cut the sequence into small fragments of length word_size, i.e., seeds. This sequence preparation method can simplify the data processing difficulty and complexity of subsequent alignment processes, and improve the accuracy and efficiency of alignment.

[0053] S300. Alignment Process: CircBLAST employs the Smith-Waterman local alignment algorithm to compare the requested sequence with sequences in the database. It considers evolutionary rearrangements of gene sequences, searching the database for partially matching subsequences and calculating the similarity and statistical significance of the matches, providing more accurate and reliable alignment results.

[0054] S400. Generating Alignment Results: CircBLAST returns alignment results in list format, including matched sequence fragments and similarity scores. These clear and concise results allow users to assess the reliability of the matches based on similarity scores and statistical significance, improving the credibility of the results. Overall, the CircBLAST tool's application method takes into account gene sequence evolution rearrangement. Through meticulous data construction, requested sequence preparation, alignment process, and result generation, it provides efficient, accurate, and reliable BLAST search functionality.

[0055] Furthermore, in S100, database construction includes the following steps:

[0056] S110. Divide protein sequences into seed blocks: Read protein sequence data from the protein database, and divide each protein sequence into seed blocks of fixed length according to the specified word_size length. The length of each seed block is word_size, and the division is done in an overlapping or non-overlapping manner.

[0057] S120. Constructing a k-mers dataset: Extract key information from protein sequence annotation data; associate each seed block with the corresponding annotation information for subsequent database construction and query operations; construct a k-mers dataset based on the associated seed blocks and annotation information.

[0058] S130. Construct a key-value data table: Design the structure of the data table, determine the keys and values ​​in the table based on the seed block and associated comment information, create a key-value data table to store the seed block and comment information, write the associated seed block and comment information into the data table, and write the data table into the database.

[0059] Specifically, segmenting protein sequences into seed blocks improves database indexing efficiency and query speed. This segmentation method breaks down large protein sequence databases into smaller, more manageable blocks, reducing computational complexity during queries and improving search efficiency. By constructing a k-mers dataset and associating seed blocks with annotation information, the relationship between protein sequences and their corresponding annotations can be more accurately linked. This association facilitates subsequent database queries and analysis, providing richer information support. Employing a key-value table structure allows for clear definition of fields and attributes, making the data structured, queryable, and manageable. Key-value database systems offer powerful query capabilities and data consistency guarantees, making data storage and access more reliable and efficient. The modular design of the construction process makes database construction and updates more flexible and scalable. This architectural design facilitates the addition or modification of functional modules to adapt to different needs and scenarios. It also provides convenience for future expansion and improvement.

[0060] Furthermore, S200 includes the following steps:

[0061] S210. Protein sequence file preparation: Obtain the protein sequence file to be compared;

[0062] S220. Sequence file format conversion: Convert protein sequence files into a format that meets the requirements;

[0063] S230, Seed Block Generation: The converted protein sequence is cut into small fragments of length word_size, i.e., seed blocks.

[0064] Specifically,

[0065] S210. Protein sequence file preparation: Obtain the protein sequence file to be compared. This can be a protein sequence file exported from a database or a file prepared by the user.

[0066] S220. Sequence File Format Conversion: Convert the protein sequence file into a suitable format, such as FASTA format. This step may include removing unnecessary annotation information and processing the file format conversion to meet the requirements of subsequent operations.

[0067] S230. Seed Block Generation: The converted protein sequence is cut into small fragments of word size, i.e., seed blocks. This step can use a fixed cutting length, such as word size, to cut the protein sequence into multiple seed blocks. Seed blocks can be divided in an overlapping or non-overlapping manner to facilitate subsequent alignment processes.

[0068] By preparing and converting protein sequence files, it can be ensured that the sequence files to be aligned conform to the format required by the program. This facilitates smooth subsequent processing and avoids data errors and format mismatches. Cutting protein sequences into seed blocks simplifies the data processing in subsequent alignment procedures. By cutting into small fragments, computational complexity can be reduced, as can the resources and time required for alignment operations. Seed block generation divides the sequence into fixed-length fragments, allowing for more precise matching of the requested sequence and fragments in the database during alignment. This helps improve the accuracy and efficiency of alignment, providing more reliable alignment results.

[0069] Furthermore, S230 includes the following steps:

[0070] S231. Traverse the sequence: For each protein sequence, traverse it starting from the beginning of the sequence according to the specified word_size;

[0071] S232, Seed block cutting: Starting from the current position, extract a segment of length word_size to form a seed block;

[0072] S233. Store seed block: Store the generated seed block in a data structure for subsequent cyclic sequence construction and comparison operations.

[0073] Specifically, in S230, further steps include:

[0074] S231. Sequence Traversal: For each protein sequence, traverse the sequence starting from the specified word_size. This step ensures a complete traversal of each sequence to generate a seed block;

[0075] S232, Seed Block Cutting: Starting from the current position, extract a fragment of length word_size to form a seed block. This involves cutting the protein sequence into small, equal-length fragments and forming a set of seed blocks.

[0076] S233. Storing Seed Blocks: The generated seed blocks are stored in a data structure for subsequent cyclic sequence construction and alignment operations. This data structure can be an array, a hash table, or other suitable data structure for storing and retrieving seed blocks. The stored seed blocks can be temporary, used for the current alignment task, or they can be part of a database for future queries and analysis.

[0077] By traversing the sequence and cutting it into seed blocks, a set containing all seed blocks is generated. This helps in identifying recurring sequence segments in subsequent cyclic sequence construction and alignment operations. Storing seed blocks and using them as identifiers for recurring sequences accelerates the alignment process in subsequent cyclic sequence construction and alignment operations. During alignment, the query sequence can be matched against the stored seed blocks, thereby reducing the number of sequence segments that need to be compared and improving alignment efficiency. Cutting the sequence into fixed-length seed blocks helps optimize the design of the alignment algorithm. Alignment operations targeting seed blocks can employ more efficient algorithms and data structures, improving the speed and accuracy of the alignment.

[0078] Furthermore, in S300, the following steps are included:

[0079] S310. Search and Match: CircBLAST calculates the index information of the hidden k-mers in each sequence using a mask. Using this index information, the requested sequence is compared with the database to find possible matching sequence fragments. The set of matching k-mers is called a Hit. If there are multiple consecutive Hits, they are called Hits.

[0080] S320. Construct circular sequences: Record the length of Hits for each sequence and the request sequence. Select the longest subsequence from all Hits as the starting point for the request sequence and the alignment sequence. Extend the selected starting sequence and the alignment sequence until the last amino acid residue before the subsequence is reached.

[0081] S330. Sequence alignment: The Smith-Waterman algorithm is used to perform sequence alignment. The alignment process considers sequence similarity, match length, and statistical significance of the match.

[0082] Specifically, the S300 includes the following steps:

[0083] S310. Search and Matching: CircBLAST uses a mask to calculate the index information of hidden k-mers in each sequence, and then uses this index information to compare the request sequence with the database to find possible matching sequence fragments. The set of matching k-mers is called a Hit, and if there are multiple consecutive Hits, they are called Hits. This step is used to filter out fragments that may be related to the request sequence to reduce the search space for subsequent comparisons.

[0084] S320. Constructing a circular sequence: In this step, the length of each sequence's Hits relative to the request sequence is recorded. The longest subsequence from all Hits is selected as the starting point for both the request and alignment sequences. The selected starting and alignment sequences are then extended until the last amino acid residue before the subsequence is reached. This constructed circular sequence contains fragments that may match the request sequence, preparing for further alignment.

[0085] S330. Sequence Alignment: Sequence alignment is performed using the Smith-Waterman algorithm. This algorithm considers sequence similarity, match length, and statistical significance of the match, evaluating the degree of sequence matching by calculating a score. Local sequence similarity is considered during the alignment process, and the best match result is returned. This step is used to precisely compare the constructed circular sequence with sequences in the database to find the optimal match.

[0086] By employing the steps of searching for matches and constructing circular sequences, segments potentially related to the requested sequence can be filtered out, and the search space for alignment can be limited, thereby improving matching accuracy. This avoids invalid alignments of irrelevant sequences, reducing noise and errors during the alignment process. Using the Smith-Waterman algorithm for sequence alignment considers sequence similarity, including local sequence similarity and match length. This helps find more accurate sequence matches and provides more reliable alignment results. During sequence alignment, a score is calculated using the Smith-Waterman algorithm to evaluate the statistical significance of the match. This score measures the confidence and significance of the match, helping to select the best matching results.

[0087] Furthermore, in S310, the index information of k-mers in each sequence is hidden by calculating the mask.

[0088] Specifically, calculating the index information of hidden k-mers in each sequence can improve the sensitivity of matching. K-mers are k consecutive bases in a sequence. By masking them, their position in the sequence can be determined. This allows for the discovery of fragments that may be related to the requested sequence, i.e., those sharing common k-mers, thus enhancing the sensitivity of matching. Because the index information of hidden k-mers is calculated, sequence fragments without common k-mers can be filtered out during alignment, reducing unnecessary alignments. This improves alignment efficiency and saves computational resources and time. Calculating the index information of hidden k-mers narrows the alignment search space to fragments that may be related to the requested sequence. This allows focus to be placed on possible matches, ignoring irrelevant sequence fragments and accelerating the search process. Since masking the index information of hidden k-mers in each sequence, sequence fragments without common k-mers can be filtered out. This helps enhance the specificity of alignment, focusing more attention on sequence fragments with high similarity.

[0089] Furthermore, in S320, the length of the Hits of each sequence and the request sequence is recorded, and the subsequence with the longest Hits is selected as the starting point of the request sequence and the alignment sequence, extending the two sequences to the last amino acid residue before the subsequence.

[0090] Furthermore, in S330, the alignment process includes the alignment of the original sequence and the circular sequence, and simultaneously includes matching fragments and similarity score information in both cases.

[0091] Furthermore, the S400 includes the following steps:

[0092] S410. Alignment result data structure: CircBLAST returns alignment results in list form, with each alignment result item including the matched sequence fragment and similarity score;

[0093] S420. Extract sequence fragments: Extract matching sequence fragments from the circular sequence obtained during the alignment process. The sequence fragments are the fragments in the matching sequence that correspond to the request sequence, describing the similarity and matching relationship between the two sequences.

[0094] S430. Calculate similarity score: Calculate a similarity score for each alignment result to measure the degree of similarity between the aligned segments;

[0095] S440. Add statistically significant information: Add statistically significant information to the comparison results. Statistically significant information includes the expected value (E-value) and confidence level of the comparison.

[0096] S450, Organizational Comparison Results: Users judge the reliability of the match based on similarity scores and statistical significance.

[0097] Specifically, S410, Alignment Result Data Structure: CircBLAST returns a list of alignment results, with each result item including the matched sequence fragment and a similarity score. This step organizes the alignment results into a data structure, facilitating subsequent processing and analysis.

[0098] S420. Sequence Fragment Extraction: Extract matching sequence fragments from the circular sequences obtained during the alignment process. Sequence fragments are segments in the matching sequence that correspond to the request sequence, describing the similarity and matching relationship between the two sequences. This step extracts the matching fragments from the alignment results for further research and analysis.

[0099] S430. Calculate the similarity score: Calculate a similarity score for each alignment result to measure the degree of similarity between the aligned segments. The similarity score can be calculated based on different algorithms and standards, typically considering factors such as sequence matching length, matching quality, and possible random matching. This step provides a quantitative indicator of the similarity of the aligned segments, helping users assess the reliability of the match.

[0100] S440. Add statistically significant information: Add statistically significant information to the comparison results, such as the expected value (E-value) and confidence level. The E-value refers to the expected number of times a result equal to or better than the current comparison result will be obtained in a random matching scenario. Confidence level refers to the reliability and significance of the comparison result. This step provides a more comprehensive evaluation metric for the comparison results;

[0101] S450. Organizing the Comparison Results: Users judge the reliability of the match based on similarity scores and statistical significance. Depending on different thresholds and strategies, users can filter and select the comparison results to obtain the most reliable matches that meet their needs. This step organizes and displays the comparison results to the user, assisting them in subsequent analysis and interpretation.

[0102] Organizing the alignment results into a data structure and extracting sequence fragments allows users to easily view and analyze each matching fragment. This facilitates further research into sequence similarity and matching relationships. Calculating similarity scores quantifies the degree of similarity between aligned fragments. This provides an objective indicator to assess the quality and reliability of the matches, facilitating result filtering and interpretation. Adding statistical significance information to the alignment results, such as E-value and confidence level, provides further evaluation and interpretation. This helps users more accurately understand the significance and reliability of the alignment results. Based on the similarity score and statistical significance information, users can filter and interpret the alignment results according to their needs and thresholds. This step empowers users with the right to interpret and control the results, increasing flexibility and customizability.

[0103] Furthermore, the application method of CircBLAST, a sequence search tool that takes into account gene sequence evolution rearrangement, was used for gene sequence search.

[0104] Implementation Case 1

[0105] The proposed method was validated using Limosilactobacillus reuteri 121 (GtfB; NCBI accession number: AAU08014.2) as the requested sequence. Sequence identity was then compared with four other software programs (Diamond, BLAST, mmseqs, and hmmer) as the final metric. First, a BLAST web server was run to obtain the protein sequence. Second, a seed index database was constructed using hash indexing technology, with each index consisting of three amino acid residues. Third, CircBLAST was run and identity was calculated for comparison with BLAST. Fourth, sequence alignment was used to identify domains, followed by a Student's t-test for matched samples. This process is illustrated as follows: Figure 3 As shown.

[0106] After calculating the identity, we performed a t-test, with a p-value of 8.9735 x 10⁻⁹⁹. Notably, sequence circularization can improve the identity between two sequences. We then selected *Pediococcus damosus* (hypothetical protein; NCBI accession number: AMV61683.1) from a database of approximately 5000 protein sequences. Initially, the identity between AAU08014.2 and AMV61683.1 was 0.5894. After sequence circularization, their identity was 0.7452, with a difference of 0.1558. The reason for the increased identity is as follows... Figure 4 As shown.

[0107] Implementation Case 2

[0108] We expanded the data to validate this improved strategy and compared it with other homology sequence search software (e.g., Diamond, MMseqs2, and HMMER). We used the uniprot dataset (approximately 200 million protein sequences) as our search data source. The comparison results of different software are shown below. Figure 5 As shown in the box plot, when the maximum values ​​are the same, CircBLAST improves the overall distribution of sequence identity compared to other algorithms, including the mean, quarter-point, and three-quarter-point. This result demonstrates the effectiveness of our method.

[0109] Based on cyclic permutations, this invention develops the CircBLAST algorithm, attempting to incorporate this biological strategy into BLAST. During the search, we first divide the sequences into seed blocks, forming an unordered dataset that can be considered a cyclic sequence dataset. Then, we find the starting sites for reading the query sequence and the alignment (sbjct) sequence to generate new sequences. Finally, we use the Smith-Waterman algorithm to align the two sequences. Using CircBLAST to analyze the GH70 glycosidase family, we found that CircBLAST significantly improves the accuracy of sequence alignment and can discover more rearranged sequences.

[0110] Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Anyone skilled in the art can make various modifications and alterations without departing from the spirit and scope of the present invention. Therefore, the scope of protection of the present invention should be determined by the claims.

Claims

1. A method for applying CircBLAST, a sequence search tool that also considers gene sequence evolution rearrangement, characterized in that, The application method of the sequence search tool CircBLAST, which takes into account gene sequence evolution rearrangement, includes the following steps: S100. Database Construction: Read protein sequence data from the protein database, segment all protein sequences according to the required word_size length, and construct a k-mers dataset by combining sequence annotation information. Construct a key-value data table based on the association relationship and write the key-value data table into the database. S200, Request Sequence Preparation: Convert the protein sequence file to be aligned into a format that meets the requirements, and then cut the sequence into small fragments of length word_size, i.e., seeds; S300, Alignment Process: The requested sequence is aligned with the sequences in the database. CircBLAST uses the Smith-Waterman local alignment algorithm to search for subsequences that partially match the requested sequence in the database. It completes three steps: search and matching, construction of circular sequences, and sequence alignment, and calculates the similarity and statistical significance of the matches. S400. Generate alignment results: CircBLAST returns the alignment results in list form, which includes matched sequence fragments and similarity scores. Statistically significant information is added to the alignment results, enabling users to assess the reliability of the match based on the similarity scores and statistical significance. In S300, the following steps are included: S310. Search and Match: CircBLAST calculates the index information of the hidden k-mers in each sequence using a mask. Using this index information, the requested sequence is compared with the database to find possible matching sequence fragments. The set of matching k-mers is called a Hit. If there are multiple consecutive Hits, they are called Hits. S320. Construct circular sequences: Record the length of Hits for each sequence and the request sequence. Select the longest subsequence from all Hits as the starting point for the request sequence and the alignment sequence. Extend the selected starting sequence and the alignment sequence until the last amino acid residue before the subsequence is reached. S330. Sequence alignment: The Smith-Waterman algorithm is used to perform sequence alignment. The alignment process considers sequence similarity, match length, and statistical significance of the match. In S330, the alignment process includes the alignment of the original sequence and the circular sequence, and simultaneously includes matching fragments and similarity score information in both cases.

2. The application method of CircBLAST, a sequence search tool that takes into account gene sequence evolution rearrangement, as described in claim 1, is characterized in that... In S100, database construction includes the following steps: S110. Divide the protein sequence into seed blocks: Divide each protein sequence into seed blocks of fixed length according to the specified word_size length. The length of each seed block is word_size, and the division can be done in an overlapping or non-overlapping manner. S120. Constructing a k-mers dataset: Extract key information from protein sequence annotation data; associate each seed block with the corresponding annotation information for subsequent database construction and query operations; construct a k-mers dataset based on the associated seed blocks and annotation information. S130. Construct a key-value data table: Design the structure of the data table, determine the keys and values ​​in the table based on the seed block and associated comment information, create a key-value data table to store the seed block and comment information, write the associated seed block and comment information into the data table, and write the data table into the database.

3. The application method of CircBLAST, a sequence search tool that takes into account gene sequence evolution rearrangement, as described in claim 1, is characterized in that... S200 includes the following steps: S210. Protein sequence file preparation: Obtain the protein sequence file to be compared; S220. Sequence file format conversion: Convert protein sequence files into a format that meets the requirements; S230, Seed Block Generation: The converted protein sequence is cut into small fragments of length word_size, i.e., seed blocks.

4. The application method of CircBLAST, a sequence search tool that takes into account gene sequence evolution rearrangement, as described in claim 3, is characterized in that... S230 includes the following steps: S231. Traverse the sequence: For each protein sequence, traverse it starting from the beginning of the sequence according to the specified word_size; S232, Seed block cutting: Starting from the current position, extract a segment of length word_size to form a seed block; S233. Store seed block: Store the generated seed block in a data structure for subsequent cyclic sequence construction and comparison operations.

5. The application method of CircBLAST, a sequence search tool that takes into account gene sequence evolution rearrangement, as described in claim 1, is characterized in that... In S400, the following steps are included: S410. Alignment result data structure: CircBLAST returns alignment results in list form, with each alignment result item including the matched sequence fragment and similarity score; S420. Extract sequence fragments: Extract matching sequence fragments from the circular sequence obtained during the alignment process. The sequence fragments are the fragments in the matching sequence that correspond to the request sequence, describing the similarity and matching relationship between the two sequences. S430. Calculate similarity score: Calculate a similarity score for each alignment result to measure the degree of similarity between the aligned segments; S440. Add information on statistical significance: The information on statistical significance includes the expected value (E-value) and confidence level of the comparison. S450, Organizational Comparison Results: Users judge the reliability of the match based on similarity scores and statistical significance.

6. The application method of CircBLAST, a sequence search tool that takes into account gene sequence evolution rearrangement, according to any one of claims 1-5, is characterized in that, The application method of CircBLAST, a sequence search tool that takes into account gene sequence evolution rearrangement as described in any one of claims 1-5, is used for gene sequence search.

Images