Hybrid Intelligence Optimization Algorithms And Parallelization For Protein Structure Prediction

Protein structure prediction is one of the hottest topics in bioinformatics. According to Anfinsen’s thermodynamic theory, the lowest energy value state of protein stand for stable protein structure. Thus an initio prediction method here is used to simulate the protein structure with calculation of the minimum potential energy value of protein function to get the most stable protein structure.In this paper, a two-dimensional AB non lattice model for simulated protein structure is used to study the protein potential function, with which the minimum value of the potential energy function of the two-dimensional AB non grid point model is used to predict the stability of the protein. This leads a continuous function optimization problem, which is the target for intelligent optimization algorithms.Particle swarm optimization algorithm is a kind of intelligent optimization algorithm which is based on the population with the ability of global optimization, which is used as the basis for the research of the algorithm in this paper. Based on study of the particle swarm simulated annealing algorithm and particle swarm tabu search algorithm, which can improve the performance of the algorithm, the paper proposes an improved optimization algorithm called particle swarm optimization, annealing and tabu search hybrid algorithm (PSOSATS). Experiments of four Fibonacci sequences and four short real protein sequences to calculate the minimum potential energy value of the protein for the simulation of protein conformation, shows the PSOSATS is capable to obtain lower potential energy value of protein in compared with other similar algorithms.Due to the huge number of amino acids in protein molecules, intelligent optimization algorithms suffer a lot in protein structure prediction with increased the computational load. So the paper also presents parallelization for algorithms based on OpenMP, MPI and CUDA. Experiments show that the parallel scheme improves significantly the efficiency of the intelligent optimization algorithm for protein structure prediction.