A method for accurate annotation of three-generation full-length transcripts

Abstract

The present invention discloses a method for accurately annotating three generations of full-length transcripts. The method comprises the following steps: obtaining corrected transcript structure information; parsing a human reference genome annotation file, extracting specific gene-transcript information as reference information; parsing the corrected transcript structure information, obtaining preliminary classification information, converting the abnormal transcripts, and obtaining normal data; annotating the normal data, obtaining specific gene-transcript annotations, and final transcript classification information; arranging the annotation information and classification information, and obtaining accurate annotations of three generations of full-length transcripts. Experiments have shown that the accuracy of annotations is significantly improved by annotating three generations of full-length transcripts using the method provided by the present invention. The present invention has important application value.

Description

Technical Field

[0001] The invention belongs to the field of bioinformatics, and in particular relates to a method for accurately annotating three-generation full-length transcripts. Background Art

[0002] Full-length transcript sequencing is a technology that uses third-generation sequencing technology to obtain the full-length sequence of mRNA. Compared with second-generation sequencing, third-generation sequencing has the advantage of high read length. Its read length can completely cover the length of most transcripts themselves, obtain complete transcript sequencing information, and avoid errors caused by splicing of short reads in second-generation sequencing. Therefore, full-length transcript sequencing has obvious advantages.

[0003] After obtaining the sequencing data, the structural data of the transcripts can be obtained through alignment software and structural analysis software, and then the transcripts can be annotated. Due to the short development time of the third-generation sequencing technology, there are few software for annotating full-length transcript sequencing data. Although existing software (such as SQANT software) can annotate transcripts, the accuracy rate is low. Summary of the invention

[0004] The goal of the present invention is to accurately annotate three generations of full-length transcripts.

[0005] The present invention firstly protects a method for accurately annotating three generations of full-length transcripts, which may include the following steps:

[0006] (1) Obtaining corrected transcript structure information;

[0007] (2) Parse the human reference genome annotation file and extract specific gene-transcript information as reference information;

[0008] (3) parsing the corrected transcript structure information obtained in step (1) to obtain preliminary classification information; performing transformation processing on the abnormal transcripts to obtain normal data;

[0009] (4) Annotating the normal data obtained in step (3) to obtain specific gene-transcript annotations and final transcript classification information;

[0010] (5) Organize annotation and classification information to obtain accurate annotations of three generations of full-length transcripts.

[0011] In the above method, in step (1), the corrected transcript structure information is obtained by correcting the structure information and original sequence information of three-generation full-length transcripts.

[0012] In the step (1), the specific steps of obtaining the corrected transcript structure information may be as follows:

[0013] (1-1) Determine the length consistency between the alignment position information in the structure information and the corresponding position in the original sequence information;

[0014] (1-2) Determine the base consistency between the alignment position sequence in the structural information and the original sequence;

[0015] (1-3) Determine the integrity and consistency of the original sequence and transcript structure;

[0016] (1-4) The consistency information in step (1-1), step (1-2) and step (1-3) is integrated to determine the structural accuracy in a weighted manner, and the abnormal regions are optimized and adjusted according to the above consistency data to obtain the corrected transcript structure information.

[0017] In the above method, in step (2), the human reference genome annotation file can be obtained from a public database.

[0018] In the above method, in step (2), the specific gene-transcript information needs to be converted into a data format before it can be used as reference information.

[0019] In the above method, in step (3), the specific steps of obtaining normal data can be as follows:

[0020] (3-1) Calculate the chain specificity, structural continuity, integrity and other data of each transcript based on its structural information, and perform weighted calculation to obtain the classification value;

[0021] (3-2) According to the classification value, the transcripts are divided into four categories: normal transcripts, fusion transcripts, structural variation transcripts and abnormal transcripts;

[0022] (3-3) Calculate the number of fusion genes and their fusion breakpoints for the fusion transcripts in step (3-2), cut the transcripts into several fragments based on the fusion breakpoints, and repeat step (3-1) for each fragment until a normal fragment classification is obtained; calculate the structural variation region and type of the structural variation transcripts in step (3-2), remove the structural variation region, and repeat step (3-1) for the remaining fragments as a whole until a normal fragment classification is obtained;

[0023] (3-4) Treat the normal transcripts in step (3-2) as a single entity and integrate them with the normal fragment classification obtained in step (3-3) to obtain normal data.

[0024] In the above method, in step (4), the specific steps of obtaining the transcript classification information may be as follows:

[0025] (4-1) Import the reference information obtained in step (2) and determine whether each fragment in the normal data has an intersection with the reference information: if so, extract the intersection annotation content; if not, classify the fragment as a new transcript;

[0026] (4-2) For the transcripts whose intersection annotations have been extracted in step (4-1), determine whether there is a single known transcript in the intersection annotation that is consistent with the transcript structure: if so, it is a known transcript and the corresponding annotation information is retained; if not, determine whether each exon region has unique annotation information; if it does not have unique annotation information, determine it to be a new transcript, and calculate the similarity coefficient between each annotated gene and the transcript, and take the highest bit as the annotation of the transcript; if it has unique annotation information, collect all uniquely annotated exons and determine their consistency; if the uniquely annotated exons are completely consistent, determine it to be a new transcript; if the uniquely annotated exons are inconsistent, determine it to be a fusion transcript; and calculate the annotation information required to be output for the relevant classification;

[0027] (4-3) The multi-fragment transcripts are combined and processed. If the annotations are inconsistent, they are judged as fusion transcripts. If the final classification of the transcripts is inconsistent, they are classified in the order of fusion transcripts > new transcripts > known transcripts; finally, annotation information and classification information are obtained.

[0028] The application of any of the above methods in accurately annotating three-generation full-length transcripts also falls within the protection scope of the present invention.

[0029] The present invention mainly provides a new method for realizing accurate annotation of full-length transcripts. The method is based on the high-precision transcript structure as a premise. Although the method itself does not require the quality of the transcript structure data, it is considered that high precision can significantly improve the annotation accuracy, and the result quality given by most comparison software and structure analysis software is still very large. The method provided by the present invention is a method for correcting transcript structure data of general quality, which is convenient for selection and use when needed. The method for correcting transcript structure data of general quality includes taking the complete information of the transcript structure as a premise, including comparison position information and original transcript sequence information. The accuracy of structural variations such as single base variation, insertion and deletion in the transcript structure information is confirmed by comparing the differences between the two in terms of length consistency, base consistency and integrity consistency. The accuracy of the transcript structure information itself is further confirmed according to the accuracy of the structural variation. If there is a difference that does not meet the standard, it is optimized and adjusted in the standard according to the above-mentioned consistency check result, until it meets the standard and is retained or judged to be completely abandoned as an error. This method of correcting transcript structure data can optimize and adjust some information that is of low quality or does not meet the standards without changing the complete and accurate information that meets the standards. It can significantly increase the proportion of low-quality information, thereby improving the accuracy of subsequent annotations.

[0030] The method provided by the present invention is also a novel method for realizing accurate annotation of full-length transcripts, including a type judgment method in a non-parameter stage and a transcript annotation method in a parameter stage.

[0031] The type judgment method in the parameter-free stage includes the following steps: parsing the input full-length transcript structure information, and preliminarily dividing it into four types: normal, fusion, structural variation, and abnormal according to structural continuity, alignment result chain specificity, etc.; analyzing the fusion breakpoints of the fusion type transcripts, and then dividing the fusion transcripts into several single normal transcript fragments based on the fusion breakpoints; analyzing the variant regions of the structural variation transcripts, determining the types of the remaining normal regions after removing the variant regions, retaining the normal fragments, and discarding the abnormal fragments; only making classification marks for the abnormal type transcripts, and not annotating them; encoding the normal fragment information of the normal transcripts and other types of processed transcripts for subsequent annotation processing.

[0032] The transcript annotation method in the reference stage includes the following steps: parsing the annotation file to obtain the structural information of each gene-transcript annotation as reference information; comparing the encoded information with the reference information to determine whether there is an annotation region corresponding to it; determining whether there is a known transcript in the annotation region that can match the target transcript; determining whether it is a fusion transcript of adjacent genes; splicing multi-fragment transcripts; and collating and outputting the final matched annotation information and structural classification information.

[0033] Experiments have shown that the three-generation full-length transcript sequencing data of the melanoma cell line COLO829 are annotated by the method provided by the present invention and the SQANT software, and the annotation accuracy is calculated. The results show that the annotation accuracy of the method provided by the present invention is 99%, and the annotation accuracy of the SQANT software is 94%. The method provided by the present invention significantly improves the annotation accuracy of the three-generation full-length transcript sequencing data. The present invention has important application value. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 Schematic diagram of the process for accurate annotation of three generations of full-length transcripts. DETAILED DESCRIPTION

[0035] The present invention is further described in detail below in conjunction with specific embodiments, and the examples provided are only for illustrating the present invention, rather than for limiting the scope of the present invention. The examples provided below can be used as a guide for further improvements by those of ordinary skill in the art, and do not constitute a limitation of the present invention in any way.

[0036] The experimental methods in the following examples, unless otherwise specified, are all conventional methods, and are performed according to the techniques or conditions described in the literature in the field or according to the product instructions. The materials, reagents, etc. used in the following examples, unless otherwise specified, can all be obtained from commercial channels.

[0037] Example 1: Establishment of a method for accurately annotating third-generation full-length transcript sequencing data

[0038] After a large number of experiments, the inventors of the present invention established a method for accurately annotating third-generation full-length transcript sequencing data.

[0039] The specific steps are as follows:

[0040] 1. Obtaining corrected transcript structure information

[0041] (1) Bioinformatics was used to analyze the third-generation full-length transcript sequencing data to obtain structural information.

[0042] (2) Input the structural information and the original sequence (i.e., the third-generation full-length transcript sequencing data) information for correction to obtain the corrected transcript structural information. The specific steps are as follows:

[0043] (1-1) Determine the length consistency between the alignment position information in the structure information and the corresponding position in the original sequence information;

[0044] (1-2) Determine the base consistency between the alignment position sequence in the structural information and the original sequence;

[0045] (1-3) Determine the integrity and consistency of the original sequence and transcript structure;

[0046] (1-4) The consistency information in step (1-1), step (1-2) and step (1-3) is integrated to determine the structural accuracy in a weighted manner, and the abnormal regions are optimized and adjusted according to the above consistency data to obtain the corrected transcript structure information.

[0047] 2. Parse the human reference genome annotation file obtained from the public database, extract the information of specific genes and transcripts, and convert it into a data format that is convenient for subsequent use as reference information.

[0048] 3. Analyze the input corrected transcript structure information to obtain preliminary classification information; transform the abnormal transcripts; and finally obtain normal data. The specific steps are as follows:

[0049] (3-1) Based on the structural information of each transcript, its chain specificity, structural continuity, integrity and other data are calculated, and the classification value is obtained by weighted calculation.

[0050] (3-2) According to the classification values, the transcripts are divided into four categories: normal transcripts, fusion transcripts, structural variation transcripts and abnormal transcripts. Abnormal transcripts are discarded.

[0051] (3-3) For the fusion transcripts in step (3-2), the number of fusion genes and their fusion breakpoints are calculated, and the transcripts are cut into several fragments based on the fusion breakpoints. Step (3-1) is repeated for each fragment until a normal fragment classification is obtained; for the structural variation transcripts in step (3-2), the structural variation region and type are calculated. After eliminating the structural variation region, step (3-1) is repeated for the remaining fragments as a whole until a normal fragment classification is obtained.

[0052] (3-4) Treat the normal transcripts in step (3-2) as a single entity and integrate them with the normal fragment classification obtained in step (3-3) to obtain normal data.

[0053] 4. Annotate the normal data obtained in step 3 to obtain specific gene-transcript annotations and final transcript classification information. The specific steps are as follows:

[0054] (4-1) Import the reference information obtained in step 2 and determine whether each fragment in the normal data has an intersection with the reference information: if so, extract the intersection annotation content; if not, classify the fragment as a new transcript.

[0055] (4-2) For the transcripts whose intersection annotations have been extracted in step (4-1), determine whether there is a single known transcript in the intersection annotation that is consistent with the transcript structure: if so, it is a known transcript and the corresponding annotation information is retained; if not, determine whether each exon region has unique annotation information; if it does not have unique annotation information, it is determined to be a new transcript, and the similarity coefficient between each annotated gene and the transcript is calculated, and the highest bit is taken as the annotation of the transcript; if it has unique annotation information, collect all uniquely annotated exons and determine their consistency. If the uniquely annotated exons are completely consistent, it is determined to be a new transcript; if the uniquely annotated exons are inconsistent, it is determined to be a fusion transcript. At the same time, calculate the annotation information required for the output of the relevant classification.

[0056] (4-3) Combine the multiple fragment transcripts, if the annotations are inconsistent, they are judged as fusion transcripts, and if the final classification of the transcripts is inconsistent, they are classified in the order of fusion transcripts > new transcripts > known transcripts. Output the annotation information and classification information.

[0057] 5. Organize annotation information and classification information to obtain accurately annotated third-generation full-length transcript sequencing data.

[0058] The schematic diagram of the process of the method for accurately annotating three-generation full-length transcripts established by the present invention is shown in Figure 1 .

[0059] Example 2: Comparison of the accuracy of annotation of third-generation full-length transcript sequencing data by the annotation method established in Example 1 and existing full-length transcript sequencing data annotation software

[0060] The third-generation full-length transcript sequencing data of this embodiment is the third-generation full-length transcript sequencing data of the melanoma cell line COLO829.

[0061] 1. The annotation method established in Example 1 was used to annotate the third-generation full-length transcript sequencing data, and the annotation accuracy was calculated.

[0062] The results show that the annotation accuracy of the annotation method established in Example 1 for the third-generation full-length transcript sequencing data is 99%.

[0063] 2. Use SQANT software to annotate the third-generation full-length transcript sequencing data and calculate the annotation accuracy.

[0064] The results showed that the annotation accuracy of SQANT software for third-generation full-length transcript sequencing data was 94%.

[0065] The above results show that the accuracy of annotation is significantly improved by using the method provided by the present invention to annotate third-generation full-length transcript sequencing data.

[0066] The present invention has been described in detail above. It will be apparent to those skilled in the art that the present invention may be implemented in a wide range under equivalent parameters, concentrations and conditions without departing from the spirit and scope of the present invention and without the need for unnecessary experimentation. Although the present invention provides specific embodiments, it should be understood that further improvements may be made to the present invention. In short, according to the principles of the present invention, this application intends to include any changes, uses or improvements to the present invention, including changes made by conventional techniques known in the art that depart from the scope disclosed in this application. Applications of some of the basic features may be made within the scope of the following appended claims.

Claims

1. A method for accurately annotating three generations of full-length transcripts, comprising the following steps: (1) Obtaining corrected transcript structure information; the specific steps for obtaining corrected transcript structure information are as follows: (1-1) Determine the length consistency between the alignment position information in the structure information and the corresponding position in the original sequence information; (1-2) Determine the base consistency between the alignment position sequence in the structural information and the original sequence; (1-3) Determine the integrity and consistency of the original sequence and transcript structure; (1-4) Integrating the consistency information in step (1-1), step (1-2) and step (1-3) to determine the structural accuracy in a weighted manner, and optimizing and adjusting the abnormal regions according to the above consistency data to obtain the corrected transcript structure information; (2) Parse the human reference genome annotation file and extract specific gene-transcript information as reference information; (3) Analyze the corrected transcript structure information obtained in step (1) to obtain preliminary classification information; transform the abnormal transcripts to obtain normal data; The specific steps to obtain normal data are as follows: (3-1) Calculate the chain specificity, structural continuity, and integrity data of each transcript based on its structural information, and perform weighted calculation to obtain the classification value; (3-2) According to the classification value, the transcripts are divided into four categories: normal transcripts, fusion transcripts, structural variation transcripts and abnormal transcripts; (3-3) For the fusion transcripts in step (3-2), calculate the number of fusion genes and their fusion breakpoints, cut the transcripts into several fragments based on the fusion breakpoints, and repeat step (3-1) for each fragment until a normal fragment classification is obtained; for the structural variation transcripts in step (3-2), calculate the structural variation region and type, remove the structural variation region, and repeat step (3-1) for the remaining fragments as a whole until a normal fragment classification is obtained; (3-4) treating the normal transcripts in step (3-2) as a single entity and integrating them with the normal fragment classification obtained in step (3-3) to obtain normal data; (4) Annotating the normal data obtained in step (3) to obtain specific gene-transcript annotations and final transcript classification information; The specific steps to obtain transcript classification information are as follows: (4-1) Import the reference information obtained in step (2) and determine whether each fragment in the normal data has an intersection with the reference information: if so, extract the intersection annotation content; if not, classify the fragment as a new transcript; (4-2) For the transcripts whose intersection annotations have been extracted in step (4-1), determine whether there is a single known transcript in the intersection annotation that is consistent with the structure of the transcript: if so, it is a known transcript and the corresponding annotation information is retained; if not, determine whether each exon region has unique annotation information; if it does not have unique annotation information, determine it to be a new transcript, and calculate the similarity coefficient between each annotated gene and the transcript, and take the highest bit as the annotation of the transcript; if it has unique annotation information, collect all uniquely annotated exons and determine their consistency; if the uniquely annotated exons are completely consistent, determine it to be a new transcript; if the uniquely annotated exons are inconsistent, determine it to be a fusion transcript; and calculate the annotation information required to be output for the relevant classification; (4-3) Combine and process the multi-fragment transcripts. If the annotations are inconsistent, they are judged as fusion transcripts. If the final classification of the transcripts is inconsistent, they are classified in the order of fusion transcripts > new transcripts > known transcripts. Finally, annotation information and classification information are obtained; (5) Organize annotation information and classification information to obtain accurate annotations of three generations of full-length transcripts.

2. The method according to claim 1, characterized in that: In the step (2), the human reference genome annotation file is obtained from a public database.

3. The method according to claim 1, characterized in that: In the step (2), the specific gene-transcript information needs to be converted into a data format before it can be used as reference information.

Images