A Self-learning Method for Protein Structure Prediction Based on Residue Contact Information

Abstract

A protein structure prediction method based on self-learning of residue contact information. First, use Robetta and RaptorX-Contact to obtain fragment library and contact map; secondly, learn residue pair distance distribution and establish information entropy in the first stage of population evolution The index reflects the degree of convergence of the population to achieve the purpose of self-learning; then, in the second stage of the final population evolution, use the learned residues to establish a scoring function for the distance distribution information, and assist the energy function to search the conformation space; finally, through clustering Get the final prediction result. The present invention provides a method for predicting protein structure. On the one hand, it can independently learn residue pair distance information and assist energy function to optimize conformational space; on the other hand, it can construct information entropy index to realize two-stage dynamic switching.

Description

technical field [0001] The invention relates to the fields of biological informatics, intelligent optimization and computer application, and in particular to a protein structure prediction method based on residue contact information self-learning. Background technique [0002] A protein is a biological macromolecule with a certain specific spatial structure formed by a polypeptide chain composed of amino acids in the form of "dehydration condensation" after twists and turns, and thus plays a specific function in an organism. The three-dimensional structure of proteins is of great importance in drug design, protein engineering, and biotechnology. At present, millions of protein sequences have been resolved, but most of the protein structures are unknown. Therefore, protein structure prediction is an important research problem. [0003] The gap between protein sequence and structure is mainly due to the rapid development of sequencing technology and relatively slow progress ...

AI Technical Summary

Problems solved by technology

However, the knowledge energy function of this method cannot effectively capture the long-range interaction between residues, and when predicting a target protein with a long sequence, switching between stages under a fixed cost is likely to cause "oversampling" or insufficient sampling to affect the prediction accuracy.

[0006] Therefore, existing multi-stage protein structure prediction methods are deficient in multi-stage sampling and prediction accuracy, and need to be improved

Images