Protein conformation sampling method based on niche local enhancement

Abstract

A protein conformation sampling method based on niche local enhancement comprises the following steps: under the framework of evolutionary computation, firstly, carrying out preliminary search on a conformation space by using an initial energy function; secondly, avoiding resampling of a previously explored region by knowledge obtained after each iteration, so as to obtain a plurality of potential low-energy regions; thirdly, constructing a scoring model based on distance, and enhancing the population in a local area; and finally, outputting the conformation with the lowest distance score in each iteration as a final prediction result. The protein conformation sampling method based on niche local enhancement is adopted, the problems that a traditional multi-mode method is low in sampling efficiency and high in calculation complexity can be solved, and therefore the overall prediction precision is improved.

Description

technical field [0001] The invention relates to the fields of bioinformatics and computer applications, in particular to a protein conformation sampling method based on niche local enhancement. Background technique [0002] Protein is the main bearer of life activities, supporting almost all functions of life, and most of the reactions in cells depend on protein. The working method and function of a protein depend on its unique three-dimensional structure. Therefore, obtaining its precise three-dimensional structure is crucial for understanding protein folding mechanism, analyzing protein function, and developing innovative drugs. In 2005, "Science" magazine also listed "Can protein folding be predicted?" as one of the 125 cutting-edge scientific issues in the 21st century. [0003] At present, biological wet experiment is the traditional method to determine the three-dimensional structure of protein, such as X-ray diffraction, nuclear magnetic resonance, cryo-electron micr...

AI Technical Summary

Problems solved by technology

[0006] In order to overcome the inaccurate energy function, insufficient sampling ability, low sampling efficiency, and low prediction accuracy of the existing protein structure prediction methods, the present invention provides a protein conformation sampling method based on local enhancement of niches. On the basis, the population is further enhanced by using a distance-based scoring model locally to accelerate its convergence, thereby obtaining a model that is more in line with the natural state structure, enhancing sampling capabilities, improving sampling efficiency, and improving overall prediction accuracy

Images