Gene family screening method based on sequence composition

Abstract

The invention relates to a gene family screening method based on sequence composition, and belongs to the field of bioinformatics and genome research. The sequences of the same gene family are a group of highly similar sequences. Not only key regions of the sequences are required to be similar, but also the whole sequences are required to be similar, however, a method based on BLAST similar sequence search and a method based on structural domain search focuses on key region search of the sequences. According to the invention, the overall characteristics of protein are measured through sequence composition, and on the basis, verification is carried out through sequence similarity and a structural domain, so that the whole and key areas of screened gene family members are very similar, and screening deviation is reduced. The invention has a very important value for research related to gene families.

Description

Technical Field

[0001] The invention relates to a gene family screening method based on sequence composition, and belongs to the field of bioinformatics and genome research. Background Art

[0002] A gene family is a group of highly similar sequences that originate from the same ancestral gene, encode similar proteins, and have similar biological functions. Gene families are of great value in research fields such as genetic engineering, molecular biology, and functional genomics.

[0003] At present, the method of screening gene families mainly includes two types: based on sequence similarity and based on domain. For a protein sequence, the influence of different regions of the sequence on its biological function is different. The BLAST software used for screening based on sequence similarity focuses on highly similar sequence regions, while the screening based on domain is screened by known sequence function regions. Both methods are based on the local region of the sequence during the screening process, and the screening results can ensure that the key regions of the sequences of family members are similar, but the sequence lengths and isoelectric points of family members are quite different. Members of the same gene family should not only ensure local similarity, but also ensure global similarity. A gene family screening method based on sequence composition provided by the present invention first ensures that the sequence is similar as a whole by sequence composition, and at the same time, it also ensures that the key regions of the sequences of gene family members are similar by sequence similarity and domain verification, thereby reducing screening bias. Summary of the invention

[0004] 1. Technical issues to be solved

[0005] The sequences of the same gene family are a group of highly similar sequences. Not only the key regions of the sequences are required to be similar, but also the overall similarity of the sequences is required. The BLAST similarity sequence search and the domain-based search method focus on the key regions of the sequences. The present invention measures the overall characteristics of the protein by sequence composition, and on this basis, verifies by sequence similarity and domain, thereby ensuring that the overall and key regions of the screened gene family members are very similar.

[0006] 2. Technical Solution

[0007] In order to achieve the above object, the main technical solutions adopted by the present invention include:

[0008] (1) Find the hidden Markov models of protein sequences and their domains of more than 10 confirmed target gene families based on literature or databases;

[0009] (2) Use software such as ProtParam (https: / / web.expasy.org / protparam / ) to calculate the proportion of each amino acid in the confirmed member sequences of the target gene family and the sequences to be confirmed, that is, the sequence composition. The order of amino acids must be consistent. For example, we require that amino acids be represented by internationally accepted single letters and arranged in alphabetical order;

[0010] (3) The Pearson correlation coefficient between the sequence composition of the confirmed member sequences of the target gene family and the sequence composition of the sequence to be confirmed was calculated using R language. If the Pearson correlation coefficient between the sequence to be confirmed and any confirmed member sequence is greater than or equal to 0.6, the sequence to be confirmed is considered a candidate sequence; otherwise, the sequence to be confirmed is considered not to be a member of the gene family.

[0011] (4) Construct a sequence library of the confirmed target gene family member sequences, and search for candidate sequences in the library using BLAST software. If any sequence in the library has a similarity greater than 60% with the candidate sequence, proceed to the subsequent steps; otherwise, the candidate sequence is considered not to be a member of the gene family.

[0012] (5) Using software such as NCBI CDD, search for the gene family domain in the candidate sequence. If the corresponding domain can be found, the candidate sequence is confirmed to be a member of the target gene family. Otherwise, the candidate sequence is considered not to be a member of the target gene family.

[0013] 3. Beneficial Effects

[0014] The present invention measures the overall characteristics of protein sequences by sequence composition and verifies the key regions of gene family member sequences. The advantage is that the overall and key regions of the gene family member sequences screened are very similar, thereby reducing screening bias. The present invention has very important value for research related to gene families. DETAILED DESCRIPTION

[0015] In order to better explain the present invention, the present invention is described in detail below by taking the screening of ZF-HD (Zinc finger-homeodomain) gene family members as an example.

[0016] (1) The 15 confirmed protein sequence members of the rice ZF-HD gene family were obtained from the PlantTFDB (http: / / planttfdb.gao-lab.org / ) database. From the literature (DOI: 10.1186 / s12870-019-1834-7), it was known that the identifier of the hidden Markov model of the ZF-HD gene family domain was "ZF-HD_dimer";

[0017] (2) The amino acid composition ratios of the 15 protein sequences of the rice ZF-HD gene family were calculated using ProtParam software (https: / / web.expasy.org / protparam / ). For example, the confirmed member sequence BGIOSGA020427_PA was composed of A alanine, C cysteine, D aspartic acid, E glutamic acid, F phenylalanine, G glycine, H histidine, I isoleucine, K lysine, L leucine, M methionine, N asparagine, P proline, and The order of amino acids is 9.34%, 2.33%, 3.11%, 5.84%, 3.50%, 9.34%, 5.45%, 0.78%, 3.11%, 3.89%, 3.50%, 1.56%, 10.89%, 10.51%, 7.39%, 9.34%, 3.11%, 4.67%, 0.78%, 1.56%. Now suppose that we determine whether the Asian cotton sequence evm.model.Ga11G1688 belongs to the ZF-HD gene family. The amino acid composition ratio of the sequence is calculated by the same method. In the same order, the composition ratios of the corresponding amino acids are: 6.30%, 2.76%, 4.72%, 7.48%, 4.33%, 9.06%, 5.12%, 1.97%, 5.12%, 5.91%, 3.94%, 4.33%, 6.30%, 9.45%, 6.30%, 5.51%, 5.12%, 3.94%, 0.79%, 1.57%;

[0018] (3) The Pearson correlation coefficient of the amino acid composition ratio of the confirmed sequence BGIOSGA020427_PA of the ZF-HD gene family and the unconfirmed sequence evm.model.Ga11G1688 was calculated by using the cor function of the R language. The results showed that the correlation coefficient was 0.81, which was greater than 0.6. Therefore, evm.model.Ga11G1688 was regarded as a candidate sequence.

[0019] (4) The 15 protein sequences of the rice ZF-HD gene family were constructed into a BLAST sequence library. The sequence to be confirmed evm.model.Ga11G1688 was searched in the library using BLAST software. It was found that the sequence similarity between the sequence to be confirmed and BGIOSGA026924-PA in the sequence library was 65%, which was greater than 60%, so the subsequent verification steps were continued;

[0020] (5) Through the search of NCBI CDD program, it was found that the 53-105 position of the sequence to be confirmed evm.model.Ga11G1688 is the structural domain "ZF-HD_dimer" of ZF-HD. Therefore, the candidate sequence evm.model.Ga11G1688 was officially confirmed as a member of the ZF-HD gene family.

Claims

1. A gene family screening method based on sequence composition, characterized in that: It includes the following steps: S1. Count the proportion of each amino acid in the protein sequence and use it as the sequence composition feature of the protein; S2. The similarity of the sequence composition between the determined gene family member sequence and the sequence to be confirmed is measured by the Pearson correlation coefficient. If the Pearson correlation coefficient between any determined gene family member sequence and the sequence to be confirmed is greater than or equal to 0.6, the sequence to be confirmed is considered to be a candidate sequence of the gene family; S3. If the BLAST sequence similarity between the candidate sequence and any identified gene family member is greater than 60%, and the candidate sequence contains a domain of the gene family, the candidate sequence is considered to be a member of the gene family.

2. A gene family screening method based on sequence composition as claimed in claim 1, characterized in that: In step S1, the amino acid composition ratio of the protein sequence is used as its feature. This feature can measure the protein sequence as a whole.

3. A gene family screening method based on sequence composition as claimed in claim 1, characterized in that: In step S2, the similarity of the sequence compositions is measured by the Pearson correlation coefficient.

4. A gene family screening method based on sequence composition as claimed in claim 1, characterized in that: In step S3, verification is performed by sequence similarity and domain structure at the same time. Both verification methods focus on the local characteristics of the sequence. In this way, the present invention takes into account both the overall characteristics and the local characteristics of the protein sequence, thereby screening more reliable gene family members.