Genome assembly correction methods, devices, and computer program products

Abstract

This application discloses a method, apparatus, and computer program product for correcting genome assembly. The method includes: correcting the initial assembly result of a sample based on long-read sequencing data to obtain a first corrected result; correcting the first corrected result based on short-read sequencing data to obtain a second corrected result and a correction record; and filtering out erroneous corrections from the second corrected result based on the short-read sequencing data and the correction record to obtain the target assembly result of the sample. This application solves the problem in related technologies where the accuracy of the assembly result is low after first correcting the initial genome assembly result with long-read data and then with short-read data.

Description

Technical Field

[0001] This application relates to the field of bioinformatics, and more specifically, to a method, apparatus, and computer program product for correcting genome assembly. Background Technology

[0002] In recent years, with the continuous deepening of genomics research and the rapid development of technology, the application of long-read sequencing in genome assembly has been greatly promoted. Compared with traditional short-read sequencing, long-read sequencing has the characteristic of longer sequencing fragments, which allows it to span repetitive regions in the genome, thereby forming telomere-to-telomere assembly results.

[0003] However, while long-read data offers significant advantages in assembly integrity and continuity, providing comprehensive and accurate genomic information, its sequencing accuracy is relatively low, typically between 90% and 99%. That is, long-read data may introduce a higher error rate during assembly, leading to lower accuracy in the assembly results.

[0004] To overcome this problem, correction tools in related technologies suggest first correcting the initial assembly results with long-read data, and then correcting them with short-read data. Examples include correction methods, devices, and computer programs such as tpolish and wtdbg2-polisher for Pilon, Racon, and Ne genome assemblies. Although these genome correction tools can generate higher quality assembly results than the original assembly, they still have some significant drawbacks.

[0005] On the one hand, due to the varying accuracy of different read lengths of data, the assembly results of some long-read data, even after correction, may still be inferior to the original assembly results of other long-read data. For example, the accuracy of ONT (Oxford Nanopore Technologies) data is lower than that of HiFi data, and the assembly results based on ONT data are still of lower quality than those based on HiFi (High Fidelity) data even after correction. On the other hand, the correction methods in related technologies are prone to over-correction, leading to result bias. That is, after multiple rounds of correction, the accuracy of the assembly results may even decrease, failing to achieve the expected results.

[0006] There is currently no effective solution to the problem that the accuracy of the assembly results is not high when the preliminary genome assembly results are corrected first with long read data and then with short read data. Summary of the Invention

[0007] This application provides a method, apparatus, and computer program product for correcting genome assembly, in order to solve the problem that the accuracy of the assembly results is not high when the preliminary assembly results of the genome are first corrected using long read data and then corrected using short read data.

[0008] According to one aspect of this application, a method for correcting genome assembly is provided. The method includes: correcting an initial assembly result of a sample based on long-read sequencing data to obtain a first corrected result; correcting the first corrected result based on short-read sequencing data to obtain a second corrected result and a correction record; and filtering erroneous corrected results from the second corrected result based on the short-read sequencing data and the correction record to obtain a target assembly result of the sample.

[0009] Optionally, correcting the initial assembly result of the sample based on the long-read sequencing data of the sample to obtain a first correction result includes: comparing the long-read sequencing data and the initial assembly result to obtain a first alignment result; identifying assembly abnormal data based on the first alignment result, and correcting the initial assembly result based on the coverage of the assembly abnormal data and the initial assembly result to obtain the first correction result.

[0010] Optionally, comparing the long-read sequencing data and the initial assembly result to obtain the first alignment result includes: extracting the minimum alignment substring from the long-read sequencing data and the initial assembly result respectively to obtain the first minimum alignment substring and the second minimum alignment substring; matching each first minimum alignment substring with the second minimum alignment substring in the initial assembly result to obtain the first matching result, wherein the first matching result includes: whether each first minimum alignment substring is successfully matched, and the position of the second minimum alignment substring matched by the successfully matched first minimum alignment substring in the initial assembly result; and determining the first alignment result based on the first matching result.

[0011] Optionally, determining the first alignment result based on the first matching result includes: connecting the first minimum aligned substring that was successfully matched in the long read sequencing data to obtain a chain; expanding seed chains in the chain based on the first minimum aligned substring that was successfully matched, and determining the initial first alignment result based on all the expanded seed chains; determining the alignment quality of the first minimum aligned substring and the second minimum aligned substring, and determining the first alignment result based on the alignment quality and the initial first alignment result.

[0012] Optionally, determining assembly aberration data based on the first alignment result and correcting the initial assembly result based on the coverage of the assembly aberration data and the initial assembly result to obtain a first correction result includes: determining a first type of assembly aberration data based on the first alignment result; correcting the first type of assembly aberration data in the initial assembly result based on long-read sequencing data to obtain a first sub-correction result, wherein the first type of assembly aberration data is assembly aberration data shorter than a preset length; analyzing the coverage of the first sub-correction result and the second type of assembly aberration data in the first sub-correction result, wherein the second type of assembly aberration data is assembly aberration data longer than or equal to a preset length; correcting the first sub-correction result if the coverage does not meet the coverage requirement to obtain a second sub-correction result with coverage meeting the coverage requirement; determining the aberration region where the second type of assembly aberration data is located in the second sub-correction result; and reassembling the aberration region based on adjacent reads and paired reads to obtain the first correction result.

[0013] Optionally, correcting the first correction result based on the short-read sequencing data of the sample to obtain the second correction result and correction record includes: comparing the short-read sequencing data and the first correction result to obtain the second alignment result; dividing the first correction result into multiple fragments, obtaining the second sub-alignment results of each fragment from the second alignment result to obtain multiple second sub-alignment results; for each fragment, constructing a partial sequence alignment graph based on the second sub-alignment results, and generating a first target consensus sequence based on the target path in the partial sequence alignment graph; and determining the second correction result and correction record based on the first target consensus sequence of the multiple fragments.

[0014] Optionally, comparing the short-read sequencing data with the first correction result to obtain the second alignment result includes: extracting the minimum alignment substring from the short-read sequencing data and the first correction result respectively to obtain the third minimum alignment substring and the fourth minimum alignment substring; matching each third minimum alignment substring with the fourth minimum alignment substring in the first correction result to obtain the second matching result, wherein the second matching result includes: whether each third minimum alignment substring is successfully matched, and the position of the fourth minimum alignment substring matched by the successfully matched third minimum alignment substring in the first correction result; and determining the second alignment result based on the second matching result.

[0015] Optionally, before dividing the first correction result into multiple segments and obtaining second sub-comparison results for each of the multiple segments from the second alignment result, the method further includes: deleting alignment sequences below a preset quality from the second alignment result to obtain an updated second alignment result; obtaining second sub-comparison results for each of the multiple segments from the second alignment result, which includes: obtaining second sub-comparison results for each of the multiple segments from the second alignment result to obtain multiple initial second sub-comparison results; calculating the base quality of the initial second sub-comparison results, and deleting initial second sub-comparison results with base quality greater than a preset value from the multiple initial second sub-comparison results to obtain multiple second sub-comparison results.

[0016] Optionally, the process of filtering out erroneous correction results from the second correction results based on short-read sequencing data and correction records to obtain the target assembly result of the sample includes: constructing a first database of sequence fragments of length k based on short-read sequencing data, and calculating the frequency of each sequence fragment of length k in the first database to obtain first frequency data; constructing a second database of sequence fragments of length k based on the first correction results, and calculating the frequency of each sequence fragment of length k in the second database to obtain second frequency data; and filtering out erroneous correction results from the second correction results based on the first and second frequency data and correction records to obtain the target assembly result.

[0017] Optionally, the process of filtering out erroneous correction results from the second correction result based on the first frequency data, the second frequency data, and the correction record to obtain the target assembly result includes: based on the correction record, listing the consensus sequences generated by each type of correction during the process of correcting the first correction result to the second correction result, thereby obtaining multiple consensus sequences; based on the first frequency data and the second frequency data, calculating the number of erroneous sequence segments and the number of missing sequence segments in each consensus sequence; determining the second target consensus sequence from the multiple consensus sequences based on the number of erroneous sequence segments and the number of missing sequence segments; and determining the target assembly result based on the second target consensus sequence.

[0018] According to another aspect of this application, a correction device for genome assembly is provided. The device includes: a first correction unit for correcting an initial assembly result of a sample based on long-read sequencing data of the sample to obtain a first correction result; a second correction unit for correcting the first correction result based on short-read sequencing data of the sample to obtain a second correction result and a correction record; and a correction unit for filtering erroneous correction results from the second correction result based on the short-read sequencing data and the correction record to obtain a target assembly result of the sample.

[0019] According to another aspect of this application, a computer program product is provided, including a non-volatile computer-readable storage medium storing a computer program that, when executed by a processor, implements a method for correcting genome assembly.

[0020] According to another aspect of this application, an electronic device is also provided, comprising a processor and a memory; the memory stores computer-readable instructions, and the processor is used to execute the computer-readable instructions, wherein the computer-readable instructions, when executed, perform a genome assembly correction method.

[0021] This application employs the following steps: First, the initial assembly result of the sample is corrected based on long-read sequencing data to obtain a first correction result; second, the first correction result is corrected based on short-read sequencing data to obtain a second correction result and a correction record; third, erroneous correction results are filtered from the second correction result based on the short-read sequencing data and the correction record to obtain the target assembly result of the sample. This solves the problem in related technologies where the accuracy of the assembly result is low after first correcting the initial genome assembly result with long-read data and then with short-read data. By alternately correcting the assembly result with long-read and short-read sequencing data, and filtering out erroneous correction results from the correction records of short-read sequencing data, the correction process does not introduce new errors, thus improving the accuracy of the assembly result. Attached Figure Description

[0022] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. In the drawings:

[0023] Figure 1 This is a flowchart of a correction method for genome assembly according to an embodiment of this application;

[0024] Figure 2 This is an optional comparison diagram of long-read sequencing data and initial assembly results according to an embodiment of this application;

[0025] Figure 3 This is a schematic diagram illustrating an optional method of correcting the initial assembly result based on the coverage of assembly anomaly data and the initial assembly result, according to an embodiment of this application.

[0026] Figure 4 This is a schematic diagram illustrating the correction of the first correction result based on short-read sequencing data of optional samples according to embodiments of this application;

[0027] Figure 5This is a schematic diagram illustrating the optional method of determining a second target consensus sequence based on the number of erroneous sequence fragments and missing sequence fragments according to an embodiment of this application;

[0028] Figure 6 This is a schematic diagram of a correction method for genome assembly according to an embodiment of this application;

[0029] Figure 7 This is a schematic diagram of the test results of chromosomes 1-22 using the ONT-R9 sequencing platform according to an embodiment of this application;

[0030] Figure 8 This is a schematic diagram of the test results of chromosomes 1-22 using the ONT-R10 sequencing platform according to an embodiment of this application;

[0031] Figure 9 This is a schematic diagram of the QV (quantity) assessment results of chromosomes 1-22 on the ONT-R9 sequencing platform according to an embodiment of this application;

[0032] Figure 10 This is a schematic diagram of the evaluation results (NG50) of chromosomes 1-22 of the ONT-R9 sequencing platform according to an embodiment of this application;

[0033] Figure 11 This is a schematic diagram of the completion results of chromosomes 1-22 on the ONT-R9 sequencing platform according to an embodiment of this application;

[0034] Figure 12 This is a schematic diagram of the QV (quantitative value) results of chromosomes 1-22 from the ONT-R10 sequencing platform according to an embodiment of this application.

[0035] Figure 13 This is a schematic diagram of the evaluation results (NG50) of chromosomes 1-22 of the ONT-R10 sequencing platform according to an embodiment of this application;

[0036] Figure 14 This is a schematic diagram of the completion results of chromosomes 1-22 on the ONT-R10 sequencing platform according to an embodiment of this application;

[0037] Figure 15 This is a schematic diagram of a correction device for genome assembly according to an embodiment of this application;

[0038] Figure 16 This is a schematic diagram of an electronic device according to an embodiment of this application. Detailed Implementation

[0039] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0040] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0041] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate for the embodiments of this application described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0042] It should be noted that all information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for display, data used for analysis, etc.) involved in this disclosure are information and data authorized by the user or fully authorized by all parties.

[0043] It should be noted that the information collected in this application is information and data authorized by the user or fully authorized by all parties, and the collection, storage, use, processing, transmission, provision, disclosure and application of the relevant data all comply with the relevant laws, regulations and standards of the relevant regions, necessary confidentiality measures have been taken, and they do not violate public order and good morals. Corresponding operation portals are provided for users to choose to authorize or refuse to use the data.

[0044] According to embodiments of this application, a method for correcting genome assembly is provided.

[0045] Figure 1 This is a flowchart of a genome assembly correction method according to an embodiment of this application. Figure 1 As shown, the method includes the following steps:

[0046] Step S102: Correct the initial assembly result of the sample based on the long read sequencing data of the sample to obtain the first correction result.

[0047] The initial assembly result of the sample refers to the longer continuous sequence obtained by recombining a large number of shorter DNA sequence fragments obtained from long-read sequencing. Long-read sequencing refers to DNA (Deoxyribonucleic acid) or RNA (ribonucleic acid) sequencing methods that can generate sequences of thousands of nucleotides or even longer.

[0048] The long-read sequencing data of the sample refers to the DNA or RNA sequence information obtained by sequencing genomic data using long-read sequencing technology. For example, the long-read sequencing data can be ONT (Oxford Nanopore Technologies) data or Cyclone SEQ-WT02 data, that is, data that determines the base sequence by detecting changes in the electrical signal as a single nucleic acid molecule passes through a nanopore.

[0049] It should be noted that long-read sequencing data has relatively low sequencing accuracy, leading to poor accuracy in the initial assembly results. For example, ONT data can generate very long reads, suitable for exploring complex and repetitive regions in the genome, but its accuracy is lower than HiFi data, especially when dealing with homopolymers and tandem repeats, where the error rate is higher. Therefore, this embodiment first uses long-read sequencing data to self-correct the initial assembly results, improving their accuracy and obtaining the first corrected result.

[0050] Step S104: Correct the first correction result based on the short read sequencing data of the sample to obtain the second correction result and correction record.

[0051] The short read sequencing data of the sample refers to a shorter DNA sequence fragment (between 50 and 300 base pairs) obtained using Next Generation Sequencing (NGS) technology.

[0052] It should be noted that, considering the high sequencing accuracy of short-read sequencing data (which can exceed 99.99%), this embodiment further corrects the first correction result of the initial assembly using short-read sequencing data, thereby effectively correcting errors in the first correction result and obtaining a second correction result. During the correction process, a correction record is generated, which includes the location of the correction and the type of correction performed.

[0053] Step S106: Based on the short read sequencing data and correction records, filter out the erroneous correction results from the second correction results to obtain the target assembly result of the sample.

[0054] It should be noted that this embodiment can perform at least two rounds of correction on the assembly results by alternating between long-read sequencing data and short-read sequencing data. Since over-correction may lead to bias during the correction process, this embodiment considers the higher sequencing accuracy of short-read sequencing data. Based on the short-read sequencing data and correction records, it determines whether there are erroneous correction results in the second correction result of each round. If so, the erroneous correction results are filtered out, ensuring that each round of correction can effectively correct errors without introducing new errors. For ONT data, the accuracy of the assembly results is further improved during multiple rounds of correction without affecting the continuity and integrity of the assembly, and without any decline in performance indicators.

[0055] The genome assembly correction method provided in this application corrects the initial assembly result of the sample based on long-read sequencing data to obtain a first correction result; it then corrects the first correction result based on short-read sequencing data to obtain a second correction result and a correction record; finally, it filters out erroneous correction results from the second correction result based on the short-read sequencing data and the correction record to obtain the target assembly result of the sample. This solves the problem in related technologies where the accuracy of the assembly result is low after first correcting the initial genome assembly result with long-read data and then with short-read data. By alternately correcting the assembly result with long-read and short-read sequencing data and filtering out erroneous correction results with the correction records of short-read sequencing data, the method achieves the effect of not introducing new errors during the correction process and improving the accuracy of the assembly result.

[0056] Since the sequence order of long-read sequencing data is scrambled, while the sequence order of the initial assembly result is known, in order to determine the position of the long-read sequencing data and correct the initial assembly result based on the long-read sequencing data, optionally, in the genome assembly correction method provided in this application embodiment, correcting the initial assembly result of the sample based on the long-read sequencing data to obtain a first correction result includes: comparing the long-read sequencing data and the initial assembly result to obtain a first alignment result; determining assembly abnormal data based on the first alignment result, and correcting the initial assembly result based on the coverage of the assembly abnormal data and the initial assembly result to obtain the first correction result.

[0057] In this embodiment, the long-read sequencing data can be the ONT fastq file, and the initial assembly result can be the preliminarily assembled FASTA file (FASTA is a text format used to represent nucleic acid or protein sequences). The ONT fastq file is used as the query sequence, and the FASTA file is used as the reference sequence. The ONT fastq file (FASTA is a text format used to store biological sequences and corresponding quality assessments) is aligned to the FASTA file to generate a BAM file (Binary Alignment / Map, a binary format for storing alignment results of high-throughput sequencing data) containing alignment information, which is the first alignment result.

[0058] Furthermore, the BAM file is scanned to find regions where the long-read sequencing data and the initial assembly results are inconsistent. Based on the inconsistent regions, abnormal assembly data are identified, and the initial assembly results are corrected based on the abnormal assembly data. The coverage of the initial assembly results is determined based on the FASTQ file, and the initial assembly results are corrected if the coverage is uneven, so that the initial assembly results have uniform coverage, thus obtaining the first correction result.

[0059] In this embodiment, abnormal assembly data are identified based on the alignment results of long-read sequencing data and initial assembly results. The initial assembly results are then corrected based on the abnormal assembly data and coverage, thereby reducing allele bias and improving the alignment accuracy in repetitive regions.

[0060] To improve alignment efficiency, optionally, in the genome assembly correction method provided in this application embodiment, aligning long-read sequencing data and initial assembly results to obtain a first alignment result includes: extracting minimum alignment substrings from the long-read sequencing data and initial assembly results respectively to obtain a first minimum alignment substring and a second minimum alignment substring; matching each first minimum alignment substring with the second minimum alignment substring in the initial assembly result to obtain a first matching result, wherein the first matching result includes: whether each first minimum alignment substring is successfully matched, and the position of the second minimum alignment substring matched by the successfully matched first minimum alignment substring in the initial assembly result; and determining the first alignment result based on the first matching result.

[0061] For example, a first minimum alignment substring can be obtained by extracting representative and unique substrings from long-read sequencing data using a weighted minimum sampler algorithm, and a second minimum alignment substring can be obtained by extracting representative and unique substrings from the initial assembly result using a minimum sampler algorithm. Then, each first minimum alignment substring is matched with the second minimum alignment substring in the initial assembly result to determine the position of the first minimum alignment substring in the initial assembly result.

[0062] Specifically, if the first minimum matching substring has a matching second minimum matching substring in the initial assembly result, the first minimum matching substring is determined to be a successful match, and the position of the matching second minimum matching substring in the initial assembly result is determined as the position of the first minimum matching substring. If the first minimum matching substring does not have a matching second minimum matching substring in the initial assembly result, the first minimum matching substring is determined to be a failed match, and the case of successful or unsuccessful matching, as well as the position determined when successful, is determined as the first matching result.

[0063] The matching process is also the comparison process. Furthermore, the first matching result can be determined as the first comparison result, or multiple first minimum matching substrings in the first matching result can be restored to complete sequences to obtain a more complete first comparison result.

[0064] Figure 2 This is an optional alignment diagram of long-read sequencing data and initial assembly results according to an embodiment of this application, such as... Figure 2 As shown, for the same sample (Ancestral genome), long read sequencing data is first obtained from the individual (Sequencedindividual) through long read sequencing, and then these long read sequencing data are assembled into a preliminary assembly result (Assembly genome, which can also be called Reference genome in this embodiment). (The reference genome, by aligning long-read sequencing data to the initial assembly, identifies sites in the reference genome that may differ from the sequencing data; these sites are called mutations. Aligning long-read sequencing data segments to the initial assembly yields alignment results. Incorrect mappings may occur during this process, indicating inconsistencies between the long-read sequencing data and the initial assembly. During alignment, the MCAS (minimal confidently alignable substrings) of the long-read sequencing data (lang reads) and the initial assembly (Reference, which acts as a reference in this alignment process) is determined. The MCAS score measures the alignment accuracy. Specifically, the reads in the long-read sequencing data are aligned to the initial assembly (Ref) to obtain the MCAS.)

[0065] This embodiment matches the minimum alignment substring in the long read sequencing data and the initial assembly results. Since the minimum alignment substring is a short substring that is representative and unique, it requires less storage and outputs faster. It can quickly find possible matching regions in a large amount of data, realize the anchoring of long read sequencing data in the initial assembly results, reduce alignment time and resource consumption, and lay the foundation for correcting the initial assembly results based on long read sequencing data.

[0066] To obtain a more complete first alignment result, optionally, in the genome assembly correction method provided in the embodiments of this application, determining the first alignment result based on the first matching result includes: connecting the first minimum alignment substring that has been successfully matched in the long read sequencing data to obtain a chain; expanding seed chains in the chain based on the first minimum alignment substring that has been successfully matched, and determining the initial first alignment result based on all the expanded seed chains; determining the alignment quality of the first minimum alignment substring and the second minimum alignment substring, and determining the first alignment result based on the alignment quality and the initial first alignment result.

[0067] For example, the first minimum alignment substring, also known as the minimizing sampler, is the process of connecting the first minimum alignment substrings (minimizing samplers) that have successfully matched in long-read sequencing data to obtain a chain. This process can be implemented using a dynamic programming algorithm (e.g., Smith-Waterman). The dynamic programming algorithm is used to identify and connect the minimizing samplers with the highest scores, effectively finding a chain with the highest score, which represents the optimal matching path in the sequence. The score of each minimizing sampler can be calculated by combining positional differences (increasing the score if the distance between two minimizing samplers is within a predefined threshold, and decreasing the score otherwise) and matching quality (adjusting the score based on the number of matched bases and mismatches; more matched bases result in a higher score, and more mismatches result in a lower score).

[0068] It should be noted that since the first minimum alignment substring only represents a portion of the long-read sequencing data, to form a more complete alignment, local alignments are performed on high-scoring strands to expand the seed strand. The initial first alignment result is determined by all expanded seed strands. This process can be implemented using a dynamic programming algorithm, which can further align adjacent sequence fragments based on the seed strand, thereby expanding and refining the alignment result. That is, the first minimum alignment substring is extended based on its position in the original long-read sequencing data. For example, if the first minimum alignment substring is 10 dp and represents a 100 dp sequencing fragment, this 10 dp first minimum alignment substring is restored to the 100 dp sequencing fragment to generate a complete initial first alignment result.

[0069] The initial first alignment result includes the matching degree of each first minimum aligned substring and the second minimum aligned substring, the position of the first minimum aligned substring in the initial assembly result, etc. After obtaining the initial first alignment result, the alignment quality is further determined based on the matching degree of each first minimum aligned substring and the second minimum aligned substring, mismatch penalty and insertion / deletion penalty mechanism (e.g., points are added for correctly matched bases, points are deducted for incorrectly matched bases, and points are deducted for inserted or missing bases). The MCAS score (minimal confidently alignable substrings) is determined based on the alignment quality, and these MCAS scores are integrated into the initial first alignment result to generate a BAM file containing alignment information of long-read sequencing data aligned to the preliminary assembly result, which is the final first alignment result.

[0070] In this embodiment, after obtaining the matching results of the smallest alignment substring in the long read sequencing data and the initial assembly results, the first successfully matched smallest alignment substring is linked into a chain, and then the local seed chain is expanded to form a complete alignment. The MCAS score is calculated, and the complete alignment result MCAS score is used to obtain the first alignment result, which can accurately and completely characterize the alignment of the long read sequencing data and the initial assembly results.

[0071] To improve the accuracy and efficiency of correction, optionally, in the genome assembly correction method provided in this application embodiment, determining assembly aberration data based on the first alignment result and correcting the initial assembly result based on the coverage of the assembly aberration data and the initial assembly result to obtain the first correction result includes: determining a first type of assembly aberration data based on the first alignment result; correcting the first type of assembly aberration data in the initial assembly result based on long-read sequencing data to obtain a first sub-correction result, wherein the first type of assembly aberration data is assembly aberration data shorter than a preset length; analyzing the coverage of the first sub-correction result and the second type of assembly aberration data of the first sub-correction result, wherein the second type of assembly aberration data is assembly aberration data longer than or equal to a preset length; correcting the first sub-correction result if the coverage does not meet the coverage requirement to obtain a second sub-correction result with coverage meeting the coverage requirement; determining the aberration region where the second type of assembly aberration data is located in the second sub-correction result; and reassembling the aberration region based on adjacent reads and paired reads to obtain the first correction result.

[0072] The preset length can be 50bp. The first alignment result is scanned to find inconsistencies between the long read sequencing data and the initial assembly result within 50bp (e.g., single nucleotide variants, insertion errors less than 50bp, deletions less than 50bp). Small errors or small variants are obtained, which are the first type of assembly anomalous data. The small errors and small variants in the initial assembly result are corrected using long read sequencing data to obtain the first sub-correction result.

[0073] Furthermore, the coverage of the first sub-correction results is analyzed. Ideally, each region should have sufficient sequencing depth (for local regions, coverage should be at least ±10x of the overall average coverage to be considered sufficient). If a region has high coverage, it indicates that the region has become narrower, possibly due to the overlap of two regions into one, resulting in high coverage. This high-coverage region can be lengthened. If a region has low coverage, it indicates that the region is very small, possibly due to incorrect initial assembly results. Fine-tuning can be performed, or the region may not exist. This region can be turned off to ensure sufficient coverage for each region and adequate sequencing depth for the entire first sub-correction result.

[0074] Furthermore, it should be noted that the original assembly results may contain gaps or incorrectly assembled regions. These gaps can be filled by reassembling using local read information, while simultaneously correcting larger assembly errors to ensure assembly accuracy. Within a local 2kb range, all reads aligned to that region can be assembled to obtain a consistent sequence. Specifically, the first sub-correction results are analyzed to identify inconsistencies of more than 50bp between long-read sequencing data and the initial assembly results (e.g., insertion errors greater than or equal to 50bp, deletions greater than or equal to 50bp), revealing potential assembly errors and significant structural variations, i.e., second-type assembly anomalies. The regions containing second-type assembly anomalies are identified in the second sub-correction results, resulting in anomaly regions. The reads flanking the anomaly regions and their paired reads are anchored, and the anomaly regions are reassembled based on these flanking reads and their paired reads, including correcting larger errors and filling in missing sequences, yielding the first correction result and the first correction record. If the initial assembly result is a FASTA file, the first calibration result is the calibrated FASTA file (referred to as FASTA①), and the first calibration record is a VCF file (referred to as VCF file①) that records the calibration process.

[0075] Figure 3 This is a schematic diagram illustrating an optional method for correcting the initial assembly result based on the coverage of assembly anomaly data and the initial assembly result, according to an embodiment of this application. Figure 3As shown, the left column represents the correction process, which can be implemented using the Pilon protocol, a tool for improving genome assembly results. The right column represents the correction results, including assembly improvement (saved in Fasta format) and variation detection (saved in VCF (Variant Call Format) format). During the correction process, alignment pileups are first evaluated to check and assess the positional information of each base in the genome sequencing data, identify and fix base errors, and identify single nucleotide polymorphisms and small insertions / deletions. Then, coverage and alignment discrepancies are scanned in the assembly improvement results to identify potential local misassemblies and larger insertions and deletions. Furthermore, reassemble across gaps and discrepant regions, attempt to fill gaps and fix local misassemblies, and attempt to build out the full sequence of larger insertions.

[0076] This embodiment utilizes long-read sequencing data to initially correct errors in the initial assembly results. Considering that shorter assembly anomalies affect coverage and are easier to correct, the shorter assembly anomalies are corrected quickly first, followed by coverage checks and corrections. While ensuring uniform coverage, the longer assembly anomalies that are difficult to correct are then corrected, ensuring that the correction of longer assembly anomalies is more reliable and improving the accuracy of the assembly results.

[0077] Optionally, in the genome assembly correction method provided in this application embodiment, correcting the first correction result based on the short read sequencing data of the sample to obtain the second correction result and correction record includes: dividing the first correction result into multiple fragments, obtaining the second sub-comparison results of the multiple fragments from the second alignment result, and obtaining multiple second sub-comparison results; for each fragment, constructing a partial sequence alignment map based on the second sub-comparison results, and generating a first target consensus sequence based on the target path in the partial sequence alignment map; and determining the second correction result and correction record based on the first target consensus sequence of the multiple fragments.

[0078] In this embodiment, the short read sequencing data can be the DNBSEQ fastq file, the first correction result is the fasta① file, the DNBSEQ fastq file is used as the query sequence, the fasta① file is used as the reference sequence, the DNBSEQ fastq file is aligned to the fasta① file, and a sam file containing the alignment information is generated, which is the second alignment result.

[0079] Furthermore, the SAM file is scanned, and high-quality sequences are identified based on the second alignment results, resulting in multiple sub-consensus sequences. Specifically, the first correction result can be divided into multiple smaller windows for local alignment and consensus sequence generation within each window. The window size is predefined, for example, 500 bases, and this size can be adjusted according to specific needs and data characteristics. A Partial Order Alignment (POA) graph is constructed based on the second sub-alignment results associated with each window. For example, a directed acyclic graph containing multiple sub-sequences is constructed to obtain the partial order alignment graph. Within each window of the partial order alignment graph, local alignment is performed using the POA graph. The POA graph integrates the alignment results of multiple sequences to generate a directed acyclic graph (DAG), where each node represents a base, and the weight of each edge is determined by the quality value of the base; the higher the quality value, the greater the weight. This means that when constructing the POA graph, the algorithm prioritizes bases with higher quality values, thereby improving the accuracy of the consensus sequence. Using a dynamic programming algorithm, the path with the highest score is found in the POA graph, and a consensus sequence for each window is generated, which is the first target consensus sequence. In this process, the first target consensus sequences generated by each window are spliced ​​together to form a complete consensus sequence, which generates a further corrected FASTA file (called FASTA②), which is the second correction result. The correction record generated during the correction process is also recorded, and a VCF file (called VCF file②), which records the correction process, is also recorded, which is the second correction record.

[0080] Figure 4This is a schematic diagram illustrating the correction of the first correction result based on short-read sequencing data of optional samples according to embodiments of this application, as shown below. Figure 4 As shown, the correction process includes:

[0081] First, the short-read sequencing data is laid out, that is, reads are arranged and connected according to their overlap relationships to form a preliminary assembly result. The short-read sequencing data is then compared with the first correction result, and the assembly result obtained in the previous correction step is segmented to facilitate analysis. Next, a partial order alignment (POA) graph is constructed based on the segmentation structure, and the construction and processing of the POA graph is accelerated using a single instruction multiple data (SIMD) instruction set to improve the efficiency of genome alignment and analysis. Further, segmentation is performed to obtain a consensus sequence. A consensus sequence refers to a base sequence common to all or most individuals at a specific location, determined by comparing the gene sequences of multiple individuals. In this embodiment, multiple sub-consensus sequences are determined based on the alignment results of short-read sequencing data and the first correction result. A POA-based diagram is then constructed based on these multiple sub-consensus sequences, and a second correction result is determined based on the POA diagram. This achieves the effect of further utilizing short-read sequencing data to improve the accuracy of the assembly results.

[0082] To improve alignment efficiency, optionally, in the genome assembly correction method provided in this application embodiment, aligning short-read sequencing data with the first correction result to obtain a second alignment result includes: extracting minimum alignment substrings from the short-read sequencing data and the first correction result respectively to obtain a third minimum alignment substring and a fourth minimum alignment substring; matching each third minimum alignment substring with the fourth minimum alignment substring in the first correction result to obtain a second matching result, wherein the second matching result includes: whether each third minimum alignment substring is successfully matched, and the position of the fourth minimum alignment substring matched by the successfully matched third minimum alignment substring in the first correction result; and determining the second alignment result based on the second matching result.

[0083] For example, a third minimum alignment substring can be obtained by extracting a representative and unique substring from short read sequencing data using the minimization sampler algorithm, and a fourth minimum alignment substring can be obtained by extracting a representative and unique substring from the first correction result using the minimization sampler algorithm. Then, each third minimum alignment substring is matched with the fourth minimum alignment substring in the first correction result to determine the position of the third minimum alignment substring in the first correction result.

[0084] Specifically, if the third minimum matching substring has a matching fourth minimum matching substring in the first correction result, the third minimum matching substring is determined to be a successful match, and the position of the matching fourth minimum matching substring in the first correction result is determined as the position of the third minimum matching substring. If the third minimum matching substring does not have a matching fourth minimum matching substring in the first correction result, the third minimum matching substring is determined to be a failed match, and the success or failure of the match and the position determined when the match is successful are determined as the second matching result.

[0085] The matching process is also the alignment process. Further, the second matching result can be determined as the second alignment result, or multiple third minimum alignment substrings in the second matching result can be restored to complete sequences to obtain a more complete second alignment result. For example, successfully matched third minimum alignment substrings in short-read sequencing data with similar positions can be connected to form a longer, high-scoring chain. Local alignment is performed within this high-scoring chain. The third minimum alignment substrings are extended based on their position in the original short-read sequencing data, expanding the seed chain to generate a complete initial second alignment result. The alignment quality is determined based on the matching degree between each third and fourth minimum alignment substring, mismatch penalties, and insertion / deletion penalties. Combined with the initial second alignment result, the final second alignment result is obtained.

[0086] This embodiment matches the short-read sequencing data with the smallest aligned substring in the first correction result. Since the smallest aligned substring is a short substring that is representative and unique, it requires less storage and outputs faster. It can quickly find possible matching regions in a large amount of data, thus anchoring the short-read sequencing data in the first correction result, reducing alignment time and resource consumption, and laying the foundation for correcting the first correction result based on the short-read sequencing data.

[0087] To improve the accuracy of the correction, optionally, in the genome assembly correction method provided in this application embodiment, before dividing the first correction result into multiple fragments and obtaining second sub-comparison results of multiple fragments from the second alignment result, the method further includes: deleting alignment sequences below a preset quality from the second alignment result to obtain an updated second alignment result; obtaining second sub-comparison results of multiple fragments from the second alignment result to obtain multiple second sub-comparison results includes: obtaining second sub-comparison results of multiple fragments from the second alignment result to obtain multiple initial second sub-comparison results; calculating the base quality of the initial second sub-comparison results, and deleting initial second sub-comparison results with a base quality greater than a preset base quality from the multiple initial second sub-comparison results to obtain multiple second sub-comparison results.

[0088] For example, the quality of each aligned sequence in the second alignment result is first obtained, and then a simple filtering is performed. Aligned sequences with quality below a preset threshold are identified as having a high error rate, while those with quality above the preset threshold are identified as having high confidence. Aligned sequences with high error rates are removed, and high confidence sequences are retained. For instance, a read is aligned to a reference genome, and the number of sites where the read does not match the reference genome is calculated, i.e., n. The length of the read is N, and the error rate is n / N. Reads with an error rate greater than 0.3 are removed.

[0089] Furthermore, the short-read sequencing sequences in the updated second alignment results are segmented (using a pre-defined window size of 500 bp) to obtain multiple sub-sequences. The quality of each sub-sequence is calculated, with an average base quality threshold of 10 for each window. If the average base quality of a sub-sequence is lower than 10, it is removed, resulting in high-quality sub-sequences. These high-quality sub-sequences form multiple sub-consensus sequences. In this embodiment, the alignment sequences in the second alignment results are filtered. The quality of the filtered sequences is calculated, and sub-sequences with quality below the predefined standard are removed, ensuring that the final consensus sequence uses only high-quality sub-sequences.

[0090] To further improve the accuracy of the assembly results, optionally, in the genome assembly correction method provided in this application embodiment, filtering out erroneous correction results from the second correction results based on short-read sequencing data and correction records to obtain the target assembly result of the sample includes: constructing a first database of sequence fragments of length k based on short-read sequencing data, and calculating the frequency of each sequence fragment of length k in the first database to obtain first frequency data; constructing a second database of sequence fragments of length k based on the first correction results, and calculating the frequency of each sequence fragment of length k in the second database to obtain second frequency data; and filtering out erroneous correction results from the second correction results based on the first frequency data, the second frequency data, and correction records to obtain the target assembly result.

[0091] For example, short-read sequencing data can be DNBSEQ data, where k can be 21, and sequence fragments of length k can be 21-mers. A 21-mer kmer database (referred to as readmer) is constructed using a fastq file of DNBSEQ with high accuracy (using sequencing data with at least Q30 (80% of bases having a quality greater than 30). Simultaneously, a second database of sequence fragments of length k is constructed based on the first correction result, where k can be 21, and sequence fragments of length k can be 21-mers. The first correction result is a fasta① file corrected using long-read data. Another 21-mer kmer database (referred to as seqmer) is constructed using fasta①.

[0092] Furthermore, the kmer frequencies in the reads and seqmers are calculated. Read length sequencing data is affected by coverage (coverage refers to the average number of times the same site is detected, e.g., 30 times). The same site may be detected multiple times. Therefore, when calculating the kmer frequencies in the reads, the number of times each kmer appears in the read is counted, and then the influence of coverage is removed to obtain the kmer frequencies in the reads, which is the first frequency data. The initial assembly results are not affected by coverage; therefore, the number of times each kmer appears in the seqmer can be counted, and this number of occurrences is determined as the kmer frequency in the seqmer, which is the second frequency data.

[0093] Furthermore, after removing the influence of coverage, the second correction result is judged to be whether the assembly collapse or assembly depression is based on whether the first frequency data and the second frequency data are consistent. If they are inconsistent, it indicates that the assembly has collapsed or the assembly has depression. The erroneous correction results are filtered from the second correction result in combination with the correction record to obtain the target assembly result.

[0094] This embodiment takes into account the high accuracy of short-read sequencing data and the accuracy of the database constructed from sequence fragments of short-read sequencing data. This embodiment constructs a first database of sequence fragments of length k based on short-read sequencing data, and filters the erroneous correction results in the second correction results according to the first database, thereby improving the accuracy of the second correction results.

[0095] Optionally, in the genome assembly correction method provided in this application embodiment, filtering out erroneous correction results from the second correction results based on the first frequency data, the second frequency data, and the correction record to obtain the target assembly result includes: based on the correction record, listing the consensus sequences generated by each type of correction in the process of correcting the first correction result to the second correction result, thereby obtaining multiple consensus sequences; based on the first frequency data and the second frequency data, calculating the number of erroneous sequence fragments and the number of missing sequence fragments in each consensus sequence; determining the second target consensus sequence from the multiple consensus sequences based on the number of erroneous sequence fragments and the number of missing sequence fragments; and determining the target assembly result based on the second target consensus sequence.

[0096] The process of correcting from the first correction result to the second correction result is the process of correcting using short-read sequencing data. The correction record records the corrections performed on each site and lists the consensus sequences generated by each correction during the short-read sequencing data correction process. For example, there are 4 corrections for 2 sites: no correction, both corrections, correction of the first site, and correction of the second site. These 4 corrections generate 4 consensus sequences.

[0097] Furthermore, the number of erroneous sequence segments (called error kmers) and missing sequence segments (called absent kmers) obtained for each consensus sequence are calculated. An error kmer is a kmer that, if the kmer generated by the sequence obtained after correction or without correction is not present in the read, is considered an error kmer. An absent kmer is a kmer that, if the kmer originally present in the read does not exist in the sequence obtained after correction or without correction, is considered an absent kmer.

[0098] It should be noted that, for each correction method, the error kmer and absent kmer calculated are used. In this embodiment, the correction method that can reduce error kmer and missing absent kmer to improve read consistency between the second correction result and the short-read sequencing data is selected. The consensus sequence obtained by this correction method is determined as the second target consensus sequence and saved to the FASTA file, which is the target assembly result. For example, if there are 10 sites, if correcting the first site produces 7 error kmers, and not correcting the first site produces 0 error kmers, then the first site is not corrected. The same applies to the other sites. Finally, a correction method is selected to generate a reliable consensus sequence.

[0099] Figure 5 This is a schematic diagram illustrating the optional method of determining the second target consensus sequence based on the number of erroneous sequence fragments and missing sequence fragments according to an embodiment of this application, as shown below. Figure 5 As shown, variant calls are performed. Variant calls refer to sites where short-read sequencing data and assembly results are inconsistent. The assembly result is the result corrected for long reads. Four paths are listed for sites in the reference sequence or assembly (in this example, the assembly result) after correction, from before and after correction. By analyzing the number of error kmers and absent kmers in each consensus path, correction paths that can reduce the number of error kmers and improve assembly and read consistency are selected. Among them, K... C K is the number of kmers in the assembly result. R This refers to the number of kmers in short-read sequencing data; if all the kmers are 0, it means they were not found. r This refers to the number of kmers obtained from short-read sequencing data after removing the influence of coverage, which can be directly compared with K. C For a direct comparison, K* is used to calculate whether there is a collapse or expansion.

[0100] It should be noted that the readmer is calculated based on the coverage of short read data and the genome size of the test sample. The kmer frequency in the readmer reflects the frequency of each kmer in the original read. The kmer frequency in the seqmer is calculated based on the corrected sequence (FASTA①) and reflects the frequency of each kmer in the corrected sequence. By comparing the kmer frequencies in the readmer and seqmer, the number of error kmers and absent kmers in each consensus sequence can be calculated. By analyzing the number of error kmers and absent kmers in each consensus sequence, corrections that can reduce error kmers and improve assembly and read consistency can be selected to generate reliable consensus sequences.

[0101] This embodiment can detect collapsed and swollen regions in the genome assembly results and filter out erroneous corrections, thereby selecting the correct corrections. This not only significantly improves the accuracy of the assembly results but also effectively reduces errors in downstream analysis, providing a more reliable data foundation for genome research.

[0102] Figure 6 This is a schematic diagram of a genome assembly correction method according to an embodiment of this application. Figure 6 As shown, the method includes:

[0103] ① Download the sequencing data of ONT (which can be an ONT Fastq file), the sequencing data of DNBSEQ (which can be a DNBSEQ Fastq file), and the reference genome (which can be a FASTA file, and the version can be hg38), and download the required analysis software.

[0104] ② The sequencing data of ONT were initially assembled using Flye (a genome assembly software) to obtain the preliminary assembly results of ONT.

[0105] ③ Align the ONT sequencing data to the initial ONT assembly results to obtain a BAM file containing alignment information. During alignment, the weighted minimum sampler algorithm and the concept of minimum reliable alignment substring can be used to align long-read sequencing data to the initial assembly data. This step performs well when dealing with homopolymers and complex tandem repeat regions.

[0106] ④ The initial assembly results were corrected using ONT sequencing data, i.e., preliminary self-correction, resulting in the self-corrected assembly result fasta1 and the corrected record vcf1. During correction, small errors and minor variations can be detected and corrected, while coverage and alignment differences are analyzed to identify potential assembly errors. This step utilizes the long read sequencing fragments of long-read data to improve the accuracy of the original assembly results.

[0107] ⑤ Align the DNBSEQ sequencing data to the self-corrected assembly result fasta1 to obtain a sam file containing alignment information. During alignment, pairwise sequence alignment algorithms and dynamic programming algorithms can be used to align short read data to the first-round corrected assembly result.

[0108] ⑥ The self-corrected assembly results were further corrected using DNBSEQ sequencing data, i.e., the second round of correction, yielding the corrected result fasta2 and the corrected record vcf2. During correction, the POA graph algorithm can be used, leveraging the high accuracy of short-read data to significantly improve the accuracy of the assembly results.

[0109] ⑦ A kmer database was constructed using DNBSEQ sequencing data and the self-corrected assembly result fasta1. Factors such as coverage and genome size were considered to filter out potential errors in the correction result fasta2, generating the filtered assembly result fasta3. During correction, a kmer-based filtering strategy was used to select correction processes that reduced erroneous kmers and improved assembly and read consistency, further optimizing the assembly results.

[0110] ⑧ Evaluate the accuracy, completeness, and continuity of the original assembly result FASTA and the corrected assembly result FASTA3.

[0111] This embodiment tests the preliminary assembly results of ONT-R9 and ONT-R10:

[0112] Figure 7 This is a schematic diagram of chromosome 1-22 test results from the ONT-R9 sequencing platform according to an embodiment of this application, as shown below. Figure 7 As shown, the evaluation metrics include uncorrected original data, first-round correction, second-round correction, and assembly results after overcorrection filtering, including NG50 (assembly continuity evaluation metric), Completeness (assembly integrity evaluation metric), and QV (assembly accuracy evaluation metric).

[0113] Figure 8 This is a schematic diagram of the test results of chromosomes 1-22 using the ONT-R10 sequencing platform according to an embodiment of this application, as shown below. Figure 8As shown, the evaluation metrics include uncorrected original data, first-round correction, second-round correction, and assembly results after overcorrection filtering, including NG50 (assembly continuity evaluation metric), Completeness (assembly integrity evaluation metric), and QV (assembly accuracy evaluation metric).

[0114] Figure 9 This is a schematic diagram of the QV (quantity, quality, and quantity) assessment results of chromosomes 1-22 from the ONT-R9 sequencing platform according to an embodiment of this application, as shown below. Figure 9 As shown, the assembly accuracy QV value of the assembly result after three rounds of correction improved significantly from ~30 to ~40 compared to the original assembly result before correction.

[0115] Figure 10 This is a schematic diagram of the evaluation results (NG50) of chromosomes 1-22 of the ONT-R9 sequencing platform according to an embodiment of this application, as shown below. Figure 10 As shown, the assembly continuity evaluation index NG50 was not different between the polished result after three rounds of correction and the original unpolished result before correction.

[0116] Figure 11 This is a schematic diagram of the completion results of chromosomes 1-22 from the ONT-R9 sequencing platform according to an embodiment of this application, as shown below. Figure 11 As shown, the assembly result after three rounds of correction (polished result) showed a slight improvement of 1% in the assembly integrity evaluation index (Completeness) compared to the original assembly result before correction (unpolished result).

[0117] Figure 12 This is a schematic diagram of the QV (quantity, quality, and quantity) assessment results of chromosomes 1-22 from the ONT-R10 sequencing platform according to an embodiment of this application, as shown below. Figure 12 As shown, after three rounds of correction, the assembly accuracy evaluation index QV value improved significantly from ~30 to ~40 compared to the original assembly result before correction.

[0118] Figure 13 This is a schematic diagram of the evaluation results (NG50) of chromosomes 1-22 of the ONT-R10 sequencing platform according to an embodiment of this application, as shown below. Figure 13As shown, the assembly continuity evaluation index NG50 was not different between the polished result after three rounds of correction and the original unpolished result before correction.

[0119] Figure 14 This is a schematic diagram of the completion results of chromosomes 1-22 from the ONT-R10 sequencing platform according to an embodiment of this application, as shown below. Figure 14 As shown, the assembly result after three rounds of correction (polished result) showed no difference in the assembly integrity evaluation index (Completeness) compared to the original assembly result before correction (unpolished result).

[0120] Taking the Consensus Quality Value (QV) of chr4 as an example, the preliminary assembly results for ONT-R9 and ONT-R10 are presented. For ONT-R9, the original QV of the preliminary assembly result was 33.0547. After the first round of correction with long read data, the QV increased to 33.2259; after the second round of correction with short read data, the QV further increased to 44.0039; finally, after a KMER-based filtering strategy, the QV of the assembly result reached 45.1007. For ONT-R10, the original QV of the preliminary assembly result was 41.3385. After the first round of correction with long read data, the QV increased to 42.8352; after the second round of correction with short read data, the QV further increased to 43.9445; finally, after filtering, the QV of the assembly result reached 45.2326.

[0121] These results demonstrate that this embodiment, by combining the advantages of long-read and short-read data, not only improves the accuracy of genome assembly but also maintains the integrity and continuity of the assembly results, enabling the generation of more reliable genome assembly results and providing strong technical support for genomics research.

[0122] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0123] This application also provides a genome assembly correction device. It should be noted that the genome assembly correction device of this application can be used to execute the genome assembly correction method provided in this application. The genome assembly correction device provided in this application is described below.

[0124] Figure 15 This is a schematic diagram of a genome assembly correction device according to an embodiment of this application. Figure 15 As shown, the device includes: a first correction unit 1502, a second correction unit 1504, and a correction and adjustment unit 1506.

[0125] The first correction unit 1502 is used to correct the initial assembly result of the sample based on the long-read sequencing data of the sample to obtain the first correction result.

[0126] The second correction unit 1504 is used to correct the first correction result based on the short read sequencing data of the sample, so as to obtain the second correction result and the correction record.

[0127] The correction unit 1506 is used to filter out erroneous correction results from the second correction results based on short-read sequencing data and correction records to obtain the target assembly result of the sample.

[0128] The genome assembly correction device provided in this application embodiment corrects the initial assembly result of the sample based on the long-read sequencing data of the sample using a first correction unit 1502 to obtain a first correction result; a second correction unit 1504 corrects the first correction result based on the short-read sequencing data of the sample to obtain a second correction result and a correction record; and a correction unit 1506 filters out erroneous correction results from the second correction result based on the short-read sequencing data and the correction record to obtain the target assembly result of the sample. This solves the problem in related technologies where the accuracy of the assembly result is not high when the initial genome assembly result is corrected first using long-read data and then using short-read data. By alternating between long-read and short-read sequencing data to correct the assembly result, and filtering out erroneous correction results using the correction records of short-read sequencing data, the device achieves the effect of not introducing new errors during the correction process and improving the accuracy of the assembly result.

[0129] Optionally, in the genome assembly correction device provided in the embodiments of this application, the first correction unit 1502 includes: a first alignment module, used to align long read sequencing data and initial assembly results to obtain a first alignment result; and a first correction module, used to determine assembly abnormal data based on the first alignment result, and to correct the initial assembly results based on the coverage of the assembly abnormal data and the initial assembly results to obtain a first correction result.

[0130] Optionally, in the genome assembly correction device provided in this application embodiment, the first alignment module includes: a first extraction submodule, used to extract minimum alignment substrings from long-read sequencing data and initial assembly results respectively, to obtain a first minimum alignment substring and a second minimum alignment substring; a first matching submodule, used to match each first minimum alignment substring with the second minimum alignment substring in the initial assembly result to obtain a first matching result, wherein the first matching result includes: whether each first minimum alignment substring is successfully matched, and the position of the second minimum alignment substring matched by the successfully matched first minimum alignment substring in the initial assembly result; and a first determination submodule, used to determine the first alignment result based on the first matching result.

[0131] Optionally, in the genome assembly correction device provided in the embodiments of this application, the first determining submodule includes: a connecting submodule, used to connect the first minimum alignment substring that has been successfully matched in the long read sequencing data to obtain a strand; an expanding submodule, used to expand seed strands in the strand based on the first minimum alignment substring that has been successfully matched, and to determine the initial first alignment result by all the expanded seed strands; and a first alignment result determining submodule, used to determine the alignment quality of the first minimum alignment substring and the second minimum alignment substring, and to determine the first alignment result based on the alignment quality and the initial first alignment result.

[0132] Optionally, in the genome assembly correction device provided in this application embodiment, the first correction module includes: a second determination submodule, used to determine a first type of assembly aberration data based on a first alignment result, and correct the first type of assembly aberration data in the initial assembly result based on long-read sequencing data to obtain a first sub-correction result, wherein the first type of assembly aberration data is assembly aberration data shorter than a preset length; an analysis submodule, used to analyze the coverage of the first sub-correction result and the second type of assembly aberration data of the first sub-correction result, wherein the second type of assembly aberration data is assembly aberration data longer than or equal to a preset length; a coverage correction submodule, used to correct the first sub-correction result when the coverage does not meet the coverage requirement, to obtain a second sub-correction result with coverage meeting the coverage requirement; and a reassembly submodule, used to determine the abnormal region where the second type of assembly aberration data is located in the second sub-correction result, and reassemble the abnormal region based on adjacent reads and paired reads to obtain the first correction result.

[0133] Optionally, in the genome assembly correction device provided in this application embodiment, the second correction unit 1504 includes: a second alignment module, used to align short read sequencing data and a first correction result to obtain a second alignment result; a second sub-alignment result determination module, used to divide the first correction result into multiple fragments, obtain second sub-alignment results of the multiple fragments from the second alignment result, and obtain multiple second sub-alignment results; a consensus sequence determination module, used to construct a partial sequence alignment map for each fragment based on the second sub-alignment results, and generate a first target consensus sequence based on the target path in the partial sequence alignment map; and a second correction module, used to determine the second correction result and correction record based on the first target consensus sequence of the multiple fragments.

[0134] Optionally, in the genome assembly correction device provided in this application embodiment, the second alignment module includes: a second extraction submodule, used to extract minimum alignment substrings from short read sequencing data and the first correction result respectively, to obtain a third minimum alignment substring and a fourth minimum alignment substring; a second matching submodule, used to match each third minimum alignment substring with the fourth minimum alignment substring in the first correction result to obtain a second matching result, wherein the second matching result includes: whether each third minimum alignment substring is successfully matched, and the position of the fourth minimum alignment substring matched by the successfully matched third minimum alignment substring in the first correction result; and a third determination submodule, used to determine the second alignment result based on the second matching result.

[0135] Optionally, in the genome assembly correction device provided in this application embodiment, the sub-consensus sequence determination module includes: a first deletion submodule, used to divide the first correction result into multiple fragments, obtain second sub-comparison results of multiple fragments respectively from the second alignment result, and before obtaining multiple second sub-comparison results, delete alignment sequences lower than a preset quality from the second alignment result to obtain an updated second alignment result; the second sub-comparison result determination module includes: a second sub-comparison result acquisition submodule, used to obtain second sub-comparison results of multiple fragments respectively from the second alignment result to obtain multiple initial second sub-comparison results; and a second deletion submodule, used to calculate the base quality of the initial second sub-comparison results, delete the initial second sub-comparison results of multiple initial second sub-comparison results that are greater than the preset base quality, and obtain multiple second sub-comparison results.

[0136] Optionally, in the genome assembly correction device provided in this application embodiment, the correction unit 1506 includes: a first frequency data determination module, used to construct a first database of sequence fragments of length k based on short read sequencing data, and calculate the frequency of each sequence fragment of length k in the first database to obtain first frequency data; a second frequency data determination module, used to construct a second database of sequence fragments of length k based on the first correction result, and calculate the frequency of each sequence fragment of length k in the second database to obtain second frequency data; and a filtering module, used to filter erroneous correction results from the second correction result based on the first frequency data, the second frequency data, and the correction record to obtain the target assembly result.

[0137] Optionally, in the genome assembly correction device provided in this application embodiment, the filtering module includes: an enumeration submodule for enumerating consensus sequences generated by each type of correction during the process of correcting a first correction result to a second correction result based on correction records, thereby obtaining multiple consensus sequences; a calculation submodule for calculating the number of erroneous sequence fragments and the number of missing sequence fragments in each consensus sequence based on first frequency data and second frequency data; and a target assembly result determination submodule for determining a second target consensus sequence from the multiple consensus sequences based on the number of erroneous sequence fragments and the number of missing sequence fragments, and determining a target assembly result based on the second target consensus sequence.

[0138] The aforementioned genome assembly correction device includes a processor and a memory. The first correction unit 1502, the second correction unit 1504, and the correction unit 1506 are all stored in the memory as program units. The processor executes the aforementioned program units stored in the memory to achieve the corresponding functions.

[0139] The processor contains a kernel, which retrieves the corresponding program units from memory. One or more kernels can be configured. By adjusting kernel parameters, the accuracy of the assembly results can be addressed in related technologies where long-read data is first used to correct the initial genome assembly, followed by short-read data.

[0140] The memory may include non-permanent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip.

[0141] This application also provides a computer storage medium for storing a program, wherein the program, when running, controls the device where the non-volatile storage medium is located to execute a genome assembly correction method.

[0142] This application also provides an electronic device. Figure 16 This is a schematic diagram of an electronic device according to an embodiment of this application. The electronic device 160 includes a processor and a memory; the memory stores computer-readable instructions, and the processor is used to execute the computer-readable instructions, wherein the computer-readable instructions, when executed, perform a genome assembly correction method. The electronic device in this document may be a server, PC, PAD, mobile phone, etc.

[0143] This application also provides a computer program product, including a non-volatile computer-readable storage medium storing a computer program, which, when executed by a processor, implements a method for correcting genome assembly.

[0144] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application 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.

[0145] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and 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.

[0146] 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.

[0147] 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.

[0148] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0149] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0150] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0151] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0152] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A method for correcting genome assembly, characterized in that, include: The initial assembly results of the samples are corrected based on the long-read sequencing data of the samples to obtain the first correction result; The first correction result is corrected based on the short read sequencing data of the sample to obtain a second correction result and a correction record; Based on the short read sequencing data and the correction record, the error correction results are filtered from the second correction results to obtain the target assembly result of the sample.

2. The method according to claim 1, characterized in that, The initial assembly results of the samples are corrected based on the long-read sequencing data of the samples, and the first correction result includes: The long-read sequencing data and the initial assembly results are compared to obtain the first comparison result; Based on the first comparison result, assembly anomaly data is determined, and the initial assembly result is corrected based on the coverage of the assembly anomaly data and the initial assembly result to obtain the first correction result.

3. The method according to claim 2, characterized in that, The long-read sequencing data and the initial assembly results are compared to obtain the first comparison result, which includes: The minimum alignment substrings are extracted from the long read sequencing data and the initial assembly results, respectively, to obtain the first minimum alignment substring and the second minimum alignment substring; Each first minimum matching substring is matched with the second minimum matching substring in the initial assembly result to obtain a first matching result, wherein the first matching result includes: whether each first minimum matching substring is successfully matched, and the position of the second minimum matching substring matched by the successfully matched first minimum matching substring in the initial assembly result; The first comparison result is determined based on the first matching result.

4. The method according to claim 3, characterized in that, The first comparison result is determined based on the first matching result, including: The first smallest matching substring successfully matched in the long-read sequencing data is connected to obtain a chain; The seed chain is expanded in the chain based on the first smallest matched substring that was successfully matched, and the initial first matching result is determined by all the expanded seed chains. Determine the alignment quality of the first minimum alignment substring and the second minimum alignment substring, and determine the first alignment result based on the alignment quality and the initial first alignment result.

5. The method according to claim 2, characterized in that, Based on the first comparison result, assembly anomaly data is determined, and the initial assembly result is corrected based on the coverage of the assembly anomaly data and the initial assembly result to obtain the first correction result, which includes: Based on the first alignment result, a first type of assembly aberration data is determined. Based on the long read sequencing data, the first type of assembly aberration data in the initial assembly result is corrected to obtain a first sub-correction result. The first type of assembly aberration data is assembly aberration data with a length less than a preset value. Analyze the coverage of the first sub-correction result and the second type of assembly anomaly data of the first sub-correction result, wherein the second type of assembly anomaly data is assembly anomaly data greater than or equal to the preset length; If the coverage does not meet the coverage requirement, the first sub-correction result is corrected to obtain a second sub-correction result whose coverage meets the coverage requirement; In the second sub-correction result, the abnormal region where the second type of assembled abnormal data is located is determined. Based on the adjacent read segments and paired read segments of the abnormal region, the abnormal region is reassembled to obtain the first correction result.

6. The method according to claim 1, characterized in that, The first correction result is corrected based on the short-read sequencing data of the sample to obtain a second correction result and correction record, including: The short read sequencing data is compared with the first correction result to obtain a second comparison result; The first correction result is divided into multiple segments, and second sub-comparison results for each of the multiple segments are obtained from the second comparison result, resulting in multiple second sub-comparison results; For each segment, a partial sequence alignment graph is constructed based on the second sub-comparison result, and a first target consensus sequence is generated based on the target path in the partial sequence alignment graph; The second correction result and the correction record are determined based on the first target consensus sequence of the plurality of fragments.

7. The method according to claim 6, characterized in that, The short-read sequencing data and the first correction result are compared to obtain a second comparison result, which includes: The minimum alignment substrings are extracted from the short read sequencing data and the first correction result, respectively, to obtain the third minimum alignment substring and the fourth minimum alignment substring; Each third minimum pair substring is matched with the fourth minimum pair substring in the first correction result to obtain a second matching result. The second matching result includes: whether each third minimum pair substring is successfully matched, and the position of the fourth minimum pair substring matched by the successfully matched third minimum pair substring in the first correction result. The second comparison result is determined based on the second matching result.

8. The method according to claim 1, characterized in that, Based on the short-read sequencing data and the correction records, the error correction results are filtered from the second correction results to obtain the target assembly results of the sample, including: Based on the short read sequencing data, a first database of sequence fragments of length k is constructed, and the frequency of each sequence fragment of length k in the first database is calculated to obtain first frequency data. Based on the first correction result, a second database of sequence fragments of length k is constructed, and the frequency of each sequence fragment of length k in the second database is calculated to obtain the second frequency data; The target assembly result is obtained by filtering out errors from the second correction result based on the first frequency data, the second frequency data, and the correction record.

9. A correction device for genome assembly, characterized in that, include: The first calibration unit is used to calibrate the initial assembly result of the sample based on the long read sequencing data of the sample to obtain the first calibration result; The second correction unit is used to correct the first correction result based on the short read sequencing data of the sample to obtain a second correction result and a correction record. The correction unit is used to filter out erroneous correction results from the second correction results based on the short read sequencing data and the correction record, so as to obtain the target assembly result of the sample.

10. A computer program product, characterized in that, The method includes a non-volatile computer-readable storage medium storing a computer program that, when executed by a processor, implements the correction method for genome assembly as described in any one of claims 1 to 8.

11. An electronic device comprising a memory and a processor, characterized in that, The memory stores a computer program, and the processor is configured to execute the correction method for genome assembly according to any one of claims 1 to 8 through the computer program.

Images