Parallel prediction methods for protein folding

Abstract

The invention relates to a parallel prediction method for protein folding, comprising the following steps: different protein structures are obtained through molecular dynamics simulation; different temperatures are set, and more protein structures with different conformations are obtained through copy exchange; a conjugated cap molecular segmentation method is used Segment the protein structure of different conformations, decompose each amino acid from the protein, and put a conjugate cap on each amino acid molecular fragment; calculate the electrostatic potential of the molecular fragment through quantum mechanics software, so as to fit each amino acid The charge of each atom in the molecular fragment and updates the charge of atoms in the molecular force field. In order to study the protein folding problem, the present invention organically combines the above-mentioned scheme considering the protein polarization effect with the method based on replica swapping, so as to overcome the two major problems of atomic interaction potential and sampling technology in dynamic simulation, in order to be able to Better results were obtained in protein folding simulations.

Description

technical field [0001] The invention relates to a parallel method for protein folding prediction, which belongs to biological information technology, computing method and computer virtual reality technology. Background technique [0002] The structure of a protein determines its function, and just knowing the genome sequence does not allow us to fully understand the function of a protein, let alone how it works. X-ray diffraction (X-ray diffraction) and nuclear magnetic resonance (nuclear magnetic resonance) are currently the most effective experimental methods for obtaining the three-dimensional structure of proteins. Difficulty, many important proteins do not currently have crystal structures. What is more critical is that the function of a protein is not only related to its static structure, but also closely related to its dynamic changes. Research ideas on the function of proteins. [0003] How to transform proteins from synthetic irregular structures to protein foldi...

AI Technical Summary

Problems solved by technology

[0009] The replica exchange molecular dynamics simulation (REMD) sampling technique is often used in the protein folding process. At present, it has achieved very good results in the conventional small peptide folding process. However, due to the limitation of computing resources, protein Due to the limitation of its own force field, the folds that can be obtained at present are very limited

In addition, the current method of temperature replica exchange also has the following problems: in the actual simulation process, the protein system is in the aqueous solution system, and to obtain a better sampling effect, due to the number of replicas required and the total number of the system The square root of the number of atoms is proportional, so because the number of water molecules is very large, the number of copies required for the simulation basically depends on the number of water molecules, so the amount of calculation required is also greatly increased

In fact, what we focus on is the conformational space of protein molecules. Based on this, we propose a new theoretical calculation model. Through this calculation method, in theory, the number of copies required is only related to the number of atoms in the protein, and not related to the number of atoms in the protein. The number of molecules is independent, so that the number of copies required is greatly reduced, and the conformational sampling effect of the protein we care about is no different from ordinary temperature copy exchange

Images