Method for predicting secondary structure of protein based on multiple evolution matrices

Abstract

The present invention discloses a method for predicting a secondary structure of the protein based on multiple evolution matrices. The method comprises: downloading a protein NR database and a BLAST program local software package to generate a position-specific scoring matrix PSSM of a given protein sequence, carrying out parameter adjustment on the PSI-BLAST program to obtain evolution matrices of different divergence degrees of the protein sequence; processing all eigenvectors in the evolution matrices to form multiple evolution matrix features; taking the multiple evolution matrix features as input of a classifier and evaluating classification accuracy to obtain an optimization model; and for the protein with an unknown structure, inputting the optimization model, and predicting the secondary structure of the protein. According to the method disclosed by the present invention, for a protein sequence, multiple matrices with different evolutionary divergence degrees are simultaneously used to express the protein sequence, so that the protein structure information is more fully expressed, the possibility of residue replacement is considered more comprehensively, accuracy for predicting the secondary structure of the protein is improved, and the encoding method is simple and effective.

Description

technical field [0001] The invention relates to the technical fields of bioinformatics and traditional protein sequence analysis, in particular to a protein secondary structure prediction method based on multiple evolution matrices. Background technique [0002] Protein is the main bearer of life activities in organisms and the basis of all life activities. Its physiological functions are not only reflected in the composition of amino acids, but also in its spatial structure. Therefore, predicting protein structure is an important task in the field of bioinformatics. Since protein secondary structure is the link between protein primary structure and tertiary structure, it is also a key step in predicting its tertiary structure from primary structure. When the accuracy rate of protein secondary structure prediction reaches 80%, the three-dimensional spatial structure of a protein molecule can be accurately predicted. It can be seen that the prediction of protein secondary s...

AI Technical Summary

Problems solved by technology

[0004] Orthogonal encoding is to use 20-bit binary numbers to uniquely represent a certain amino acid, and satisfy the requirement that the orthogonal product of the encoding vector values ​​of different amino acids is 0. Although the encoding method is simple, but because it carries less biological information, it makes the protein secondary The accuracy of structure prediction is low; the Codon codon coding method "reduces" the amino acid into a composition of 3 bases, and the base is represented by a binary number, thereby realizing structure prediction; The relative probability of an amino acid type appearing at a position can carry biological evolution information to a certain extent

[0005] At present, the existing traditional protein structure prediction methods generally only consider the proportion of various amino acids in the protein sequence, and there are shortcomings: this method is relatively simple, but it does not take into account the position information of amino acids in the protein and the protein evolution process Acceptable point mutations in amino acids that occur in , lacking representation of biological evolution information

[0006] In summary, in the prior art, only amino acid composition is considered when predicting the secondary structure of amino acid residues in protein sequences, and the position information of amino acids in proteins and the acceptable point mutations of amino acids that occur during protein evolution cannot be fully considered; classification However, there is still a lack of effective solutions for problems such as difficult parameter selection and poor reliability.

Images