求解非格点模型的蛋白质结构预测问题的拟物拟人算法  

Quasi-Physical and Quasi-Human Algorithm for Protein Structure Prediction in an Off-Lattice Model

作　　者：刘景发[1] 黄文奇[1] 

机构地区：[1]华中科技大学计算机学院

出　　处：《小型微型计算机系统》2006年第10期1889-1892,共4页Journal of Chinese Computer Systems

基　　金：国家"九七三"项目(2004CB318000)资助;国家自然科学基金项目(10471051)资助.

摘　　要：本文研究了一个具有两种氨基酸(疏水氨基酸和亲水氨基酸)的三维非格点的蛋白质模型.受物理世界的物体间相互作用的规律和人类社会生活经验的启发,给出了该模型蛋白质结构预测问题的拟物拟人算法.计算结果表明被提出的方法在非格点的蛋白质模型上是有效的.与文献中给出的所有算例的结果相比,无论是在最低能量值还是在计算时间上,本文算法都要好.对于这些算例中规模最大的3个,还找到了与文献中结构完全不同的最低能量构形.A three-dimensional off-lattice protein model is studied. The model has two species of different amino acids, hydrophobic and hydrophilic. Enlightened by the law of reciprocity among things in the physics world and human society, a related quasi-physical and quasi-human algorithm is formulated. Experimental results demonstrate that the proposed method is quite effective in solving the off-lattice protein structure prediction problem. Compared with the results of all instances in literatures, the proposed algorithm can get better results with respect to not only the lowest energy but also runtime. For three of these instances ,the algorithm finds new lowest energy configurations different from those in references.

关 键 词：蛋白质结构预测 三维非格点模型 拟物拟人算法 NP难度问题 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]