Local enhancement differential evolution protein conformational space searching method

Abstract

The invention discloses a local enhancement differential evolution protein conformational space searching method. The method comprises the following steps: giving an input sequence, and setting system parameters including a population size, the number of iterations, a crossed factor and a fragment length; performing complete fragment assembly on each individual in a population to generate an initial population; updating the population by executing variation, crossover and selection operation on each individual in the initial population in sequence to obtain an updated population; performing local enhancement on each individual in the updated population by calling a Monte Carlo method, and receiving enhanced individuals according to a set Boltzmann receiving probability to obtain an enhanced population; and iteratively running the above steps to reach an end condition. Through adoption of the local enhancement differential evolution protein conformational space searching method, the conformational space searching dimensions are effectively reduced; the convergence speed of an algorithm is increased; the prediction accuracy is effectively increased; and a conformational space can be sampled more effectively.

Description

technical field [0001] The invention relates to the fields of bioinformatics and computer applications, in particular to a locally enhanced differential evolution protein conformation space search method. Background technique [0002] Protein molecules play a crucial role in the chemical reactions of biological cells. Their structural models and bioactive states have important implications for our understanding and cure of many diseases. Only when proteins are folded into specific three-dimensional structures can they produce their unique biological functions. Therefore, to understand the function of a protein, it is necessary to obtain its three-dimensional spatial structure. [0003] The problem of protein structure prediction has been attracting attention since the 1950s, especially the optimization method of ab initio prediction of conformational space, which is a hot research topic in the fields of bioinformatics and computational biology, because the three-dimensiona...

AI Technical Summary

Problems solved by technology

[0005] In order to overcome the deficiencies of the existing protein structure prediction methods, such as high conformation space search dimension, slow convergence speed, and low prediction accuracy, the present invention proposes a locally enhanced differential evolution protein conformation space search method based on the differential evolution algorithm. LEDE: Under the framework of the differential evolution algorithm, the Rosetta Score3 coarse-grained knowledge energy model is used; the introduction of knowledge-based fragment assembly technology can effectively improve the prediction accuracy; the good local search performance of the Monte Carlo algorithm is used to locally enhance the population to obtain more For an excellent local conformation, combined with the strong global search ability of the differential evolution algorithm, the conformation space can be sampled more effectively

Images