基于并行群体模拟退火算法的蛋白质折叠研究

蛋白质折叠研究对于揭示蛋白结构和功能关系，进而了解相关疾病的致病机理意义重大。蛋白质折叠已被证明是 NP-完全问题。本文针对蛋白质折叠研究中的能量最小化问题，提出了一种新的并行群体模拟退火算法（Parallel Group Simulated Annealing，PGSA）及其改进型算法（PGSA_1/K)。该算法使用了降温因子加速收敛精度，并采用 MPI 消息传递并行编程技术加快蛋白质结构空间搜索以及能量最小化寻找速度。以 Met_Enkephalin 蛋白为对象的计算机模拟仿真结果表明，我们提出的算法及其改进型有很好的扩展性，可以高效搜索蛋白结构空间，从而找到相关蛋白的最小能量结构。

Protein Folding Study Based on Parallel Group Annealing Algorithms

Protein folding study plays a significant role in revealing the relationship between protein structure and function, and in understanding the pathogenesis of the related diseases. Protein folding has been shown to be a NP-complete problem. We propose a new parallel group simulated annealing algorithm (PGSA) and its variant (PGSA_1/K) to find the minimum energy structure for proteins. The algorithms use the cooling factor to improve the convergence accuracy, and employ the message passing interface (MPI) technique to accelerate the search speed of minimum energy and its configurations. The simulation results of Met_Enkephalin protein showed that our proposed algorithm and its modification have good robustness and can reduce the search space efficiently.