Research On Efficient Monte Carlo Simulation For Protein Folding

Protein engineering is the development frontier of modern biotechnology. Its fundamental purpose is to transform the naturally occurring proteins according to the idea of the people, or to design some non-natural new proteins with certain special functions. And one of the important foundations of this transformation and design is protein structure prediction uud protein folding simulation. Ths shapes cf prctein fold structure determine its biological function, and it means that there exists consistence protein structures and protein functions. Therefore, the prediction and research of protein folding has very important significance in the field of protein engineering, biological medicine, etc.In this paper, with stochastic processes, stochastic optimization and other related knowledge, we can analyze these theories of the simulated annealing algorithm and the Wang-Landau Monte Carlo methods. Then we design some more efficient parallel simulated annealing algorithms to search for the global minimum energy of the protein and its configuration. And the use of mathematical statistics, Our aim is to get an estimate of the global minimum energy, It ultimately obtains the energy range of the protein folding research. Finally, within the energy range of protein systems, we can also design some efficient parallel Wang-Landau Monte Carlo method for protein folding. And we achieve the simulation that randomly walks in this energy space. At the same time, we can obtain the state density function of the system and other information. Furthermore, we can study the entire process of protein folding thermodynamics.This paper study efficient Monte Carlo simulation methods for protein folding, main characteristics are summarized as follows:(1). We propose two efficient parallel simulated annealing algorithms--the parallel group simulated annealing algorithm (PGSA) and the new parallel simulated annealing algorithm (NPSA):Because the protein-energy space is unknown (it is difficult to obtain low energy), and combined with relevant theoretical analysis, we propose two efficient parallel simulated annealing algorithm (PGSA and NPSA) to search minimum energy and protein conformation. Then we can determine the energy range of the protein systems.(2).We design three efficient parallel Wang-Landau Monte Carlo methods:Because conventional Wang-Landau algorithm has the slow convergence problem for protein system in the low-energy range, and combined with the theoretical analysis, this paper presents three efficient parallel Wang-Landau Monte Carlo methods, they can guarantee the density of states (DOS) compared with the premise of improving the accuracy of the convergence rate.