A method for detecting fusion genes based on DNA-based sequencing data

Abstract

The application relates to the technical field of gene detection, and discloses a method for detecting fusion genes based on DNA-based sequencing data, which comprises the following steps: step S1, acquiring DNA sequencing data, filtering the sequencing data by using software fastp, and obtaining filtered sequencing data; step S2, detecting by using software Genefuse, obtaining a first fusion gene result and related parameters; step S3, detecting by using software Factera, obtaining a second fusion gene result and related parameters; step S4, detecting by using software Arriba, obtaining a third fusion gene result and related parameters; and step S5, acquiring a fusion gene filtering model, taking the first fusion gene and related parameters, the second fusion gene and related parameters, and the third fusion gene and related parameters as input data to calculate a true positive fusion gene, and outputting a result. The application constructs a fusion gene screening standard based on multiple software, avoids false positives or false negatives of a single software detection result, and improves the detection rate and correctness of the fusion gene.

Description

Technical Field

[0001] This invention relates to the field of gene detection technology, and more specifically to a method for detecting fusion genes based on DNA-based sequencing data. Background Technology

[0002] Fusion genes are new genes formed when all or part of the coding or non-coding regions of two or more different genes fuse together due to some mechanism (such as genomic variation). This is caused by chromosomal rearrangement. Fusion genes are closely related to the occurrence and development of tumors. Identifying relevant fusion genes can help identify diagnostic biomarkers, discover new therapeutic targets, and understand the molecular basis of tumorigenesis.

[0003] Currently, commonly used clinical methods for detecting fusion genes include immunohistochemistry, FISH, PCR, and next-generation sequencing (NGS). NGS technology offers a wide detection range and high speed, enabling simultaneous verification of gene fusion at both the gene and transcriptome levels, and accurately detecting breakpoints in fusion genes. This overcomes the limitations of conventional methods, such as missed detections and the inability to identify fusion partner genes. NGS methods for detecting fusion genes include DNA-based and RNA-based NGS methods.

[0004] DNA-based methods for detecting fusion genes primarily involve aligning paired-end sequencing sequences with the genome to assess whether the distance and orientation of the paired-end sequences match the library construction information, thereby determining whether it is a fusion gene. Commonly used software includes Genefuse and Factera. RNA-based methods for detecting fusion genes include sequence alignment-based methods and splicing-based methods. The former identifies fusion events by looking for inconsistent sequences and sequences covering breakpoints, while the latter assembles transcripts and aligns them with a reference genome to identify fusion transcripts consistent with chromosomal rearrangements. Commonly used software includes Arriba.

[0005] Genefuse is a DNA-based sequencing tool that can directly detect fusion genes from FASTQ files. It identifies support reads that map well to the left and right halves of two different genes but not completely to the entire reference genome, and analyzes each support read to determine if it's a fusion gene. However, this software can only focus on clinically significant genes and has limited ability to detect unknown fusion genes. Factera is a software tool for discovering fusion genes from DNA sequencing, primarily used to detect translocations, inversions, and deletions. This software first aligns and processes the original FASTQ file to obtain a BAM file, then uses the BAM file as input. It clusters near-paired exons into different genomes by finding mismatched reads, identifying breakpoints, and locating fusion genes. However, this software relies on the BAM file, is not sensitive enough to the detection results, and lacks visualization capabilities for fusion detection. Arriba is a fusion gene detection tool using RNA-based sequencing data. It can detect inversions and duplications of fusion genes quickly and sensitively, but it struggles to detect deleted fusion genes. Therefore, it is evident that each software has its drawbacks in detecting fusion genes, and using a single software to detect fusion genes may result in false positives or false negatives. Summary of the Invention

[0006] To address the above issues, this invention proposes a method for detecting fusion genes based on DNA-based sequencing data. This method constructs a screening criterion based on multiple software programs and uses mainstream current fusion gene detection software for filtering and screening, effectively avoiding false positives or false negatives from single software detection results.

[0007] A method for detecting fusion genes based on DNA-based sequencing data, comprising the following steps:

[0008] Step S1: Obtain DNA sequencing data, use the software FastP to filter the data, and obtain the filtered sequencing data;

[0009] Step S2: Use the software Genefuse to detect fusion genes in the filtered sequencing data to obtain the first fusion gene result and the first related parameter information;

[0010] Step S3: Use the software Factera to detect fusion genes in the filtered sequencing data to obtain the second fusion gene results and second related parameter information;

[0011] Step S4: Use the software Arriba to detect fusion genes in the filtered sequencing data to obtain the third fusion gene results and third related parameter information;

[0012] Step S5: Obtain the fusion gene filtering model, and use the first fusion gene result and the first related parameter, the second fusion gene result and the second related parameter, and the third fusion gene result and the third related parameter as input data. Calculate the true positive fusion genes through the fusion gene filtering model and output the fusion gene results.

[0013] Preferably, the first relevant parameter information mentioned in step S2 includes the number of breakpoints and the base quality of the breakpoint sequence, wherein the number of breakpoints is divided into total number and unique number, and the base quality is high quality, medium quality, low quality and very low quality, respectively.

[0014] Preferably, the second relevant parameter information mentioned in step S3 includes break_support, break_depth, proper_pair_support and total_depth, wherein break_support is divided into break_support1 and break_support2.

[0015] Preferably, the third relevant parameter information mentioned in step S4 includes split_read, coverage, and confidence, wherein split_read is divided into split_read1 and split_read2, coverage is divided into coverage1 and coverage2, and confidence is divided into low, medium, and high.

[0016] Preferably, the fusion gene filtering model step S5 is as follows:

[0017] Step S51: Obtain the first fusion gene result and first related parameter information, the second fusion gene result and second related parameter information, and the third fusion gene result and third related parameter information respectively, and obtain a potential fusion gene table based on the above fusion gene results and related parameter information. The potential fusion gene table includes: fusion gene pair, fusion gene 1, fusion gene 2, detection software, sequence information and related parameters.

[0018] Step S52: If the fusion gene pair is detected in two or more fusion gene detection software programs, then set the fusion gene pair as a fusion gene list and obtain the relevant parameter information;

[0019] Step S53: If a fusion gene pair is detected in only one fusion gene detection software, while its corresponding fusion gene 1 or fusion gene 2 is detected in multiple fusion gene detection software, then extract the corresponding fusion gene pair sequence information from multiple software programs, obtain abnormal alignment sequences based on DNA sequencing data, and use the BWA software to align the abnormal alignment sequences with the reference genome. Obtain the fusion gene result based on the alignment quality. If the alignment results of multiple abnormal alignment sequences are consistent, then obtain the fusion gene pair based on the alignment results, set the fusion gene pair as a fusion gene list, and obtain relevant parameter information. If the alignment results of multiple abnormal alignment sequences are inconsistent, then select the alignment result with higher quality based on the alignment quality, obtain the fusion gene pair, set the fusion gene pair as a fusion gene list, and obtain relevant parameter information. If the alignment results of multiple abnormal alignment sequences are inconsistent but the alignment quality is consistent, then select the result with higher quality based on the base quality, obtain the fusion gene pair, set the fusion gene pair as a fusion gene list, and obtain relevant parameter information.

[0020] Step S54: If the fusion gene pair is detected only in the first fusion gene result, and fusion gene 1 and fusion gene 2 are not detected in other fusion gene pairs, then obtain the first fusion gene result and related parameters for judgment: if the total number of breakpoints is ≥20, then the fusion gene pair is considered a true positive, set as a fusion gene list, and obtain related parameter information; if 10 ≤ the total number of breakpoints < 20 and the number of low-quality and very low-quality bases is less than 15% of the total length of the fusion gene pair, then extract the abnormal alignment sequence .fastq file from the DNA sequencing data according to the fusion gene pair sequence, use samtools software and picard software to align and remove duplicates from the abnormal alignment sequence .fastq file, obtain the abnormal alignment sequence .bam file, use Factera software and Arriba software respectively to detect fusion genes from the abnormal alignment sequence .bam file and abnormal alignment .fastq file, if the same fusion gene pair is detected in one of the software, then the fusion gene pair is considered a true positive, set as a fusion gene list, and obtain related parameter information;

[0021] Step S55: If the fusion gene pair is detected only in the second fusion gene result, and fusion gene 1 and fusion gene 2 are not detected in other fusion gene pairs, then obtain the relevant parameters of the second fusion gene result for judgment: if break_support ≥ 20 and both break_depth and proper_pair_support are greater than 100, then the fusion gene pair is considered a true positive, set as a fusion gene list, and obtain the relevant parameter information; if 10 ≤ break_support < 20, and break_depth, proper_pair_support, and total_depth are all greater than 100, then extract the sequence information of the fusion gene pair, extract the abnormal alignment sequence .fastq file from the DNA sequencing data, and use Genefuse software and Arriba software to detect fusion genes from the abnormal alignment sequence .fastq file. If the same fusion gene pair is detected in one of the software, then the fusion gene pair is considered a true positive, and obtain the relevant parameter information;

[0022] Step S56: If the fusion gene pair is detected only in the third fusion gene result, and fusion gene 1 and fusion gene 2 are not detected in other fusion gene results, then obtain the relevant parameters of the third fusion gene result for judgment: if confidence is high, and split_read ≥ 20 and coverage ≥ 200, then the fusion gene pair is considered a true positive, set as a fusion gene list, and obtain parameter information; if confidence is medium or low, 10 ≤ split_read < 20, and coverage ≥ 200, extract the sequence information of the fusion gene pair, and extract the abnormal alignment sequence .fastq file from the DNA sequencing data based on this information. Use BWA and picard software to analyze the abnormal alignment sequence .fastq file to obtain the abnormal alignment sequence .bam file. Use Genefuse and Factera software to detect fusion genes in the abnormal alignment sequence .fastq file and the abnormal alignment sequence .bam file, respectively. If either software detects the same fusion gene pair, then the fusion gene pair is considered a true positive, set as a fusion gene list, and obtain relevant parameter information.

[0023] Step S57: Obtain the list of fusion genes and output the results.

[0024] Compared with the prior art, the present invention has the following beneficial effects:

[0025] (1) This invention uses three software programs (Genefuse, Factera and Arriba) based on different detection algorithms and different detection data to detect fusion genes, which can make up for the detection differences of different fusion gene types by a single software. Furthermore, it incorporates an RNA-based software to detect fusion genes from DNA-based sequencing data, which can effectively increase the detection rate of fusion genes in different algorithms.

[0026] (2) This invention sets the filtering threshold of fusion genes by setting the effective parameters of each software and constructs a fusion gene filtering model, which improves the accuracy of fusion genes detected by each software and can provide more effective data support for guiding clinical diagnosis and medication. Attached Figure Description

[0027] Figure 1 This is a flowchart of a fusion gene detection method based on DNA-based sequencing data in an embodiment of the present invention.

[0028] Figure 2 This is a flowchart of the fusion gene filtering model in an embodiment of the present invention. Detailed Implementation

[0029] The present invention will now be described in detail with reference to the embodiments.

[0030] Example 1:

[0031] A method for detecting fusion genes based on DNA-based sequencing data, such as Figure 1 As shown, the specific steps are as follows:

[0032] Step S1: Obtain DNA sequencing data and use the software FastP to filter the data to obtain the filtered sequencing data.

[0033] Specifically, the data filtering criteria using the software FastP are: filtering if the base quality is less than Q20, the sequence contains N bases, or the sequence length is less than 50bp. The command line is: -q 20 -n 0 --length_required 50.

[0034] Step S2: Use the software Genefuse to detect fusion genes in the filtered sequencing data to obtain the first fusion gene result and the first related parameter information;

[0035] In the Genefuse software, the human genome hg19 is used as the reference gene for comparison, and a database of all clinically relevant fusion genes based on the reference genome hg19 is used as the fusion gene reference file.

[0036] The relevant parameters in the Genefuse software include: the number of breakpoints (including the total number and the unique number), the diff value, and the base quality of the breakpoint sequence. The total number is the total number of breakpoints detected in the fusion gene, the unique number is the number of non-repeating breakpoints detected in the fusion gene, and the diff value is expressed as (diff1|diff2), which is the edit distance between fusion gene 1 and fusion gene 2 and the reference genome, respectively. The base quality of the breakpoint sequence refers to the quality value of each base, which are: very high quality (base quality > Q40), high quality (base quality: Q30~Q39), medium quality (base quality: Q20~Q29), low quality (base quality: Q15~Q19), and very low quality (base quality: 0~Q14).

[0037] In this embodiment, the more breakpoints detected, the more reliable the breakpoints are, and the higher the quality of each base site of the fusion gene sequence, the higher the quality of the sequence and the more reliable the detection results are.

[0038] Step S3: Use the software Factera to detect fusion genes in the filtered sequencing data to obtain the second fusion gene results and second related parameter information;

[0039] In this embodiment, the steps for detecting fusion genes using the software Factera are as follows: (1) The filtered sequencing data is compared with the reference genome using the software BWA to obtain the aligned bam file; (2) The duplicates in the aligned bam file are identified and filtered using the software picard to obtain the filtered bam file; (3) The fusion gene is detected in the filtered bam file using the software Factera; wherein, the human genome hg19 and its exons are used as the reference genome and reference exons.

[0040] Specifically, the relevant parameters of the software Factera include: break_support, break_depth, proper_pair_support, unmapped_support, improper_pair_support, paired_end_support, and total_depth; among them, break_support is the number of reads that support the breakpoint, divided into break_support1 and break_support2; break_depth is the number of read pairs that cross the breakpoint; proper_pair_support: the number of correctly paired and soft-sheared read pairs; unmapped_support: the number of read pairs that were not previously mapped to the fusion; improper_pair_support: the number of inconsistent read pairs; paired_end_support: the number of read pairs flanking the breakpoint; and total_depth: the average total depth of the flanking regions of the two breakpoints.

[0041] In this embodiment, the more break_support, break_depth, and proper_pair_support there are, the more reliable the breakpoint is. The higher the total_depth is, the higher the sequencing depth near the breakpoint is, and the more reliable the breakpoint is.

[0042] Step S4: Use the software Arriba to detect fusion genes in the filtered sequencing data to obtain the third fusion gene results and third related parameter information;

[0043] In this embodiment, the software Arriba uses the human genome hs37d5 as the reference genome; relevant parameters include: split_read, discordant_mates, coverage, and confidence; where split_read is the number of read pairs supporting the breakpoint, divided into split_read1 and split_read2; discordant_mates represents the number of inconsistent read pairs supporting the fusion; coverage represents the number of fragments near the breakpoint of the fusion gene, divided into coverage1 and coverage2; confidence is the confidence level of each fusion gene, divided into low, medium, and high.

[0044] Among them, the higher the parameters split_read and coverage, the higher the confidence of the fusion gene. In this embodiment, split_read, coverage and confidence are selected as the third relevant parameters.

[0045] Step S5: Obtain the fusion gene filtering model, and use the first fusion gene result and the first related parameter, the second fusion gene result and the second related parameter, and the third fusion gene result and the third related parameter as input data. Calculate the true positive fusion genes through the fusion gene filtering model and output the fusion gene results.

[0046] The flowchart of the fusion gene filtering model is as follows: Figure 2 As shown, the specific steps include:

[0047] Step S51: Obtain the first fusion gene result and first related parameter information, the second fusion gene result and second related parameter information, and the third fusion gene result and third related parameter information respectively, and obtain a potential fusion gene table based on the above fusion gene results and related parameter information. The potential fusion gene table includes: fusion gene pair, fusion gene 1, fusion gene 2, detection software, sequence information and related parameters.

[0048] Step S52: If the fusion gene pair is detected in two or more fusion gene detection software programs, then set the fusion gene pair as a fusion gene list and obtain relevant parameter information;

[0049] Step S53: If a fusion gene pair is detected in only one fusion gene detection software, while its corresponding fusion gene 1 or fusion gene 2 is detected in multiple fusion gene detection software, then extract the corresponding fusion gene pair sequence information from multiple software programs, obtain abnormal alignment sequences based on DNA sequencing data, and use the BWA software to align the abnormal alignment sequences with the reference genome. Based on the alignment quality, obtain the fusion gene results: If the alignment results of multiple abnormal alignment sequences are consistent, obtain the fusion gene pair based on the alignment results, set the fusion gene pair as a fusion gene list, and obtain relevant parameter information; if the alignment results of multiple abnormal alignment sequences are inconsistent, select the alignment result with higher quality based on the alignment quality, set the fusion gene pair as a fusion gene list, and obtain relevant parameter information; if the alignment results of multiple abnormal alignment sequences are inconsistent but the alignment quality is consistent, select the result with higher quality based on the base quality, set the fusion gene pair as a fusion gene list, and obtain relevant parameter information.

[0050] Step S54: If the fusion gene pair is detected only in the first fusion gene result, and fusion gene 1 and fusion gene 2 are not detected in other fusion gene pairs, then obtain the first fusion gene result and related parameters for judgment: if the total number of breakpoints is ≥20, then the fusion gene pair is considered a true positive, set as a fusion gene list, and obtain related parameter information; if 10 ≤ the total number of breakpoints < 20 and the number of low-quality and very low-quality bases is less than 15% of the total length of the fusion gene pair, then extract the abnormal alignment sequence .fastq file from the DNA sequencing data according to the fusion gene pair sequence, use samtools software and picard software to align and remove duplicates from the abnormal alignment sequence .fastq file to obtain the abnormal alignment sequence .bam file, use Factera software and Arriba software respectively to detect fusion genes from the abnormal alignment sequence .bam file and the abnormal alignment sequence .fastq file, if the same fusion gene pair is detected in one of the software, then the fusion gene pair is considered a true positive, set as a fusion gene list, and obtain related parameter information;

[0051] Step S55: If the fusion gene pair is detected only in the second fusion gene result, and fusion gene 1 and fusion gene 2 are not detected in other fusion gene pairs, then obtain the relevant parameters of the second fusion gene result for judgment: if break_support ≥ 20 and both break_depth and proper_pair_support are greater than 100, then the fusion gene pair is considered a true positive, set as a fusion gene list, and obtain the relevant parameter information; if 10 ≤ break_support < 20, and break_depth, proper_pair_support, and total_depth are all greater than 100, then extract the sequence information of the fusion gene pair, extract the abnormal alignment sequence .fastq file from the DNA sequencing data, and use Genefuse software and Arriba software to detect fusion genes from the abnormal alignment sequence .fastq file. If the same fusion gene pair is detected in one of the software, then the fusion gene pair is considered a true positive, and obtain the relevant parameter information;

[0052] Step S56: If the fusion gene pair is detected only in the third fusion gene result, and fusion gene 1 and fusion gene 2 are not detected in other fusion gene results, then obtain the relevant parameters of the third fusion gene result for judgment: if confidence is high, and split_read ≥ 20 and coverage ≥ 200, then the fusion gene pair is considered a true positive, set as a fusion gene list, and obtain parameter information; if confidence is medium or low, 10 ≤ split_read < 20, and coverage ≥ 200, extract the sequence information of the fusion gene pair, and extract the abnormal alignment sequence .fastq file from the DNA sequencing data based on this information. Use BWA and picard software to analyze the abnormal alignment sequence .fastq file to obtain the abnormal alignment sequence .bam file. Use Genefuse and Factera software to detect fusion genes in the abnormal alignment sequence .fastq file and abnormal alignment sequence .bam file, respectively. If either software detects the same fusion gene pair, then the fusion gene pair is considered a true positive, set as a fusion gene list, and obtain relevant parameter information.

[0053] The steps for extracting abnormal alignment sequences are as follows: the obtained potential breakpoint sequence information is compared with the filtered sequencing data, and sequences with an alignment consistency rate greater than 95% are extracted as abnormal alignment sequences.

[0054] Step S57: Obtain the list of fusion genes and output the results.

[0055] Example 2:

[0056] According to the steps of Example 1, fusion gene detection was performed on 6 samples, and the results of the first fusion gene obtained in step S2 for the 6 samples are shown in Table 1.

[0057] Table 1: Detection results and related parameter information of fusion genes based on Genefuse software:

[0058] Sample number Fusion genes Gene 1 Gene 2 total number Unique quantity Base mass percentage of breakpoint sequences S1 EML4-ALK EML4 ALK 66 35 3.8％ S2 BCR-ALK BCR ALK 6 2 5.6％ S3 RET-KIF5B RET KIF5B 10 8 11.3％ S4 CD74-ALK CD74 ALK 12 2 2.7％ S4 NTRK1-ROS1 NTRK1 ROS1 25 2 3.1％ S5 RET-CCDC6 RET CCDC6 35 24 7.1％ S6 RET-KIF5B RET KIF5B 60 27 2.1％

[0059] In this embodiment, the base mass percentage of the breakpoint sequence is less than the base mass percentage of Q20.

[0060] According to step S3, the fusion gene was detected in the six samples, and the results of the second fusion gene were obtained, as shown in Table 2:

[0061] Table 2: Detection results and related parameter information of fusion genes based on Factera software:

[0062]

[0063]

[0064] In this context, "-" indicates an unknown fusion gene type.

[0065] According to step S4, the fusion gene was detected in the six samples, and the results of the third fusion gene were obtained, as shown in Table 3.

[0066] Table 3: Detection results and related parameter information of fusion genes based on Arriba software:

[0067]

[0068]

[0069] In this embodiment, no fusion gene was detected in sample S4 in the Arriba software.

[0070] Based on step S5, the fusion gene filtering model was used to calculate the first, second, and third fusion gene results for the six samples to obtain the true positive fusion genes. The fusion gene results are shown in Table 4.

[0071] Table 4: Final fusion gene results:

[0072] Sample number Fusion genes Software 1 Software 2 Software 3 Detection status S1 RRM1-AS1-IQGAP3 × × √ × S1 EML4-ALK √ √ √ √ S2 ASXL1-RP5-963E22.4 × × √ × S2 BCR-ALK √ × × × S2 FGF7-ANKRD20A2 × √ × √ S2 PAIP2B-NUTM1 × √ × × S2 YTHDF1-ASXL1 × √ × × S3 RET-KIF5B √ √ × √ S3 HLA-DRB5-HLA-DRB1 × √ × √ S3 LYZL2-RET × √ √ √ S4 NTRK1-ROS1 √ × × √ S4 TUBA3D-GNAQ × √ × × S5 RET-CCDC6 √ √ × √ S5 RET-TCERG1L × × √ × S6 RET-KIF5B √ × × √ S6 SLC36A4-PNPLA3 × √ √ √ S6 CCL11-PNPLA3 × √ × × S6 PNPLA3-TWIST2 × √ × × S6 RET-RASGEF1A × √ × √ S6 SUMF1-PNPLA3 × × √ × S6 RRM1-AS1-IQGAP3 × × √ ×

[0073] Based on the above results, it is evident that using the method of this invention to detect fusion genes increases the accuracy of fusion gene detection compared to using a single software. For example, the fusion gene LYZL2-RET in sample S3 was detected in both Factera and Arriba software, but at a low frequency. Using a single software might filter out the results of this fusion gene. However, by using the method of this invention, the detection results from two software programs are cross-validated, which increases the accuracy of fusion gene detection. Secondly, this invention can effectively reduce false negatives in fusion gene detection. For example, the fusion gene FGF7-ANKRD20A2 in sample S2 was detected only in Factera software, and at a high frequency. After calculation by the fusion gene filtering model, this fusion gene was determined to be a true positive. Therefore, this invention can reduce false positives and false negatives in fusion gene detection based on DNA-based sequencing data, and improve the accuracy of fusion gene detection.

Claims

1. A method for detecting fusion genes based on DNA-based sequencing data for non-disease diagnostic purposes, characterized by the following steps: include: Step S1: Obtain DNA sequencing data, use the software FastP to filter the data, and obtain the filtered sequencing data; Step S2: Use the software Genefuse to detect fusion genes in the filtered sequencing data to obtain the first fusion gene result and the first related parameter information; Step S3: Use the software Factera to detect fusion genes in the filtered sequencing data to obtain the second fusion gene results and second related parameter information; Step S4: Use the software Arriba to detect fusion genes in the filtered sequencing data to obtain the third fusion gene results and third related parameter information; Step S5: Obtain the fusion gene filtering model, and use the first fusion gene result and the first related parameter, the second fusion gene result and the second related parameter, and the third fusion gene result and the third related parameter as input data. Calculate the true positive fusion genes through the fusion gene filtering model and output the fusion gene results. The steps of the fusion gene filtering model described in step S5 are as follows: Step S51: Obtain the first fusion gene result and first related parameter information, the second fusion gene result and second related parameter information, and the third fusion gene result and third related parameter information respectively, and obtain a potential fusion gene table based on the above fusion gene results and related parameter information. The potential fusion gene table includes: fusion gene pair, fusion gene 1, fusion gene 2, detection software, sequence information and related parameters. Step S52: If the fusion gene pair is detected in two or more fusion gene detection software programs, then set the fusion gene pair as a fusion gene list and obtain relevant parameter information; Step S53: If a fusion gene pair is detected in only one fusion gene detection software, while its corresponding fusion gene 1 or fusion gene 2 is detected in multiple fusion gene detection software, then extract the corresponding fusion gene pair sequence information from multiple software programs, obtain abnormal alignment sequences based on DNA sequencing data, and use the BWA software to align the abnormal alignment sequences with the reference genome. Based on the alignment quality, obtain the fusion gene results: If the alignment results of multiple abnormal alignment sequences are consistent, obtain the fusion gene pair based on the alignment results, set the fusion gene pair as a fusion gene list, and obtain relevant parameter information; if the alignment results of multiple abnormal alignment sequences are inconsistent, select the alignment result with higher quality based on the alignment quality, set the fusion gene pair as a fusion gene list, and obtain relevant parameter information; if the alignment results of multiple abnormal alignment sequences are inconsistent but the alignment quality is consistent, select the result with higher quality based on the base quality, set the fusion gene pair as a fusion gene list, and obtain relevant parameter information. Step S54: If the fusion gene pair is detected only in the first fusion gene result, and fusion gene 1 and fusion gene 2 are not detected in other fusion gene pairs, then obtain the first fusion gene result and related parameters for judgment: if the total number of breakpoints is ≥20, then the fusion gene pair is considered a true positive, set as a fusion gene list, and obtain related parameter information; if 10 ≤ the total number of breakpoints < 20 and the number of low-quality and very low-quality bases is less than 15% of the total length of the fusion gene pair, then extract the abnormal alignment sequence .fastq file from the DNA sequencing data according to the fusion gene pair sequence, use samtools software and picard software to align and remove duplicates from the abnormal alignment sequence .fastq file to obtain the abnormal alignment sequence .bam file, use Factera software and Arriba software respectively to detect fusion genes from the abnormal alignment sequence .bam file and the abnormal alignment sequence .fastq file, if the same fusion gene pair is detected in one of the software, then the fusion gene pair is considered a true positive, set as a fusion gene list, and obtain related parameter information; Step S55: If the fusion gene pair is detected only in the second fusion gene result, and fusion gene 1 and fusion gene 2 are not detected in other fusion gene pairs, then obtain the relevant parameters of the second fusion gene result for judgment: if break_support ≥ 20 and both break_depth and proper_pair_support are greater than 100, then the fusion gene pair is considered a true positive, set as a fusion gene list, and obtain the relevant parameter information; if 10 ≤ break_support < 20, and break_depth, proper_pair_support, and total_depth are all greater than 100, then extract the sequence information of the fusion gene pair, extract the abnormal alignment sequence .fastq file from the DNA sequencing data, and use Genefuse software and Arriba software to detect fusion genes from the abnormal alignment sequence .fastq file. If the same fusion gene pair is detected in one of the software, then the fusion gene pair is considered a true positive, and obtain the relevant parameter information; Step S56: If the fusion gene pair is detected only in the third fusion gene result, and fusion gene 1 and fusion gene 2 are not detected in other fusion gene results, then obtain the relevant parameters of the third fusion gene result for judgment: if confidence is high, and split_read ≥ 20 and coverage ≥ 200, then the fusion gene pair is considered a true positive, set as a fusion gene list, and obtain parameter information; if confidence is medium or low, 10 ≤ split_read < 20, and coverage ≥ 200, extract the sequence information of the fusion gene pair, and extract the abnormal alignment sequence .fastq file from the DNA sequencing data based on this information. Use BWA and picard software to analyze the abnormal alignment sequence .fastq file to obtain the abnormal alignment sequence .bam file. Use Genefuse and Factera software to detect fusion genes in the abnormal alignment sequence .fastq file and abnormal alignment sequence .bam file, respectively. If either software detects the same fusion gene pair, then the fusion gene pair is considered a true positive, set as a fusion gene list, and obtain relevant parameter information. Step S57: Obtain the list of fusion genes and output the results.

2. The method for detecting fusion genes based on DNA-based sequencing data according to claim 1, characterized in that, The first relevant parameter information mentioned in step S2 includes the number of breakpoints and the base quality of the breakpoint sequence. The number of breakpoints is divided into total number and unique number, and the base quality is divided into extremely high quality, high quality, medium quality, low quality and extremely low quality.

3. The method for detecting fusion genes based on DNA-based sequencing data according to claim 1, characterized in that, The second related parameter information mentioned in step S3 includes break_support, break_depth, proper_pair_support, and total_depth, where break_support is divided into break_support1 and break_support2.

4. The method for detecting fusion genes based on DNA-based sequencing data according to claim 1, characterized in that, The third relevant parameter information mentioned in step S4 includes split_read, coverage, and confidence, where split_read is divided into split_read1 and split_read2, coverage is divided into coverage1 and coverage2, and confidence is divided into low, medium, and high.

5. The method for detecting fusion genes based on DNA-based sequencing data according to claim 1, characterized in that, In steps S53-S56, the step of extracting the abnormal alignment sequence is as follows: the obtained potential breakpoint sequence information is compared with the filtered sequencing data, and sequences with an alignment consistency rate greater than 95% are extracted as abnormal alignment sequences.

Images