Multi-objective optimization protein structure prediction method based on residue contact diagram

Abstract

The invention discloses a multi-objective optimization protein structure prediction method based on a residue contact diagram. The method comprises the following steps: firstly, predicting the residuecontact diagram of a target protein sequence by utilizing Triplet Res, MetaPSICOV, RaptorX and STOPConact; secondly, designing a scoring function to initialize the population; then, sampling the population through fragment recombination and assembly; and finally, for different prediction servers, calculating according to a designed conact energy function to obtain four conact energies, sorting all conformations in the population through a weighted scoring function Etotal (Cn) which harmonizes an average number and a standard deviation, and selecting a first conformation in the last generationas a prediction result. According to the invention, protein misfolding caused by inaccurate prediction of a single residue contact pattern can be relieved, so that the diversity is increased, and theoverall prediction precision is improved.

Description

technical field [0001] The invention relates to the fields of bioinformatics and computer applications, in particular to a multi-objective optimization protein structure prediction method based on a residue contact map. Background technique [0002] Protein is the cornerstone of life, almost all cellular activities involve protein, and the three-dimensional structure of protein determines its special biological function. Therefore, protein structure information is crucial in protein research. For example, the catalytic function of an enzyme is performed by a part of the protein chain, that is, the active site exposed on the surface of the protein. Interactions between proteins and between proteins and molecules of nucleic acids, inhibitors, and activators are also limited to specific protein surface areas. Therefore, it is only possible to design targeted drugs that interact with protein surfaces if the protein structure is known. At present, the protein structure is main...

AI Technical Summary

Problems solved by technology

[0005] In order to alleviate the problem of insufficient accuracy of protein structure prediction caused by inaccurate contact information of a single residue pair, the present invention provides a multi-objective optimized protein structure prediction method based on the residue contact map. Under the framework of the basic genetic algorithm, a contact scoring function is designed , according to the residue contact maps predicted by the four contact prediction servers, the conformations are scored, sorted and selected, thereby increasing the population diversity and improving the overall prediction accuracy

Images