一种用于多重生物序列比对的协同进化算法  

CO-EVOLUTIONARY ALGORITHM FOR MULTIPLE BIOLOGICAL SEQUENCE ALIGNMENT

作　　者：张树波[1,2] 

机构地区：[1]中山大学数学与计算科学学院,广东广州510275 [2]广东省信息安全重点实验室,广东广州510275

出　　处：《计算机应用与软件》2010年第8期37-40,共4页Computer Applications and Software

基　　金：国家自然科学基金资助(60675016;60633030)

摘　　要：序列比对是生物序列分析的一项基础性工作,多重生物序列比对可以帮助预测未知序列的结构和功能。在传统遗传算法的基础上,提出了一种用于多重生物序列比对的协同进化算法,通过引进协同进化机制,使得保守区域和位点以较高的概率传给子代,而非保守区域和位点以较低的概率传给子代。在BAliBASE3.0数据库上,该方法获得比常用的比对算法ClustralX更高的适应度,比传统的遗传算法具有更好的收敛性。Alignment of biological sequence is the basic work for biological sequence analysis. Multiple sequence alignment can help to predict the structure and function of unknown sequences. A co-evolutionary algorithm is proposed for muhiple sequence alignment based on traditional genetic algorithm. By introducing co-evolutionary mechanism, the conserved regions and loci are able to be transmitted to offspring in higher probability, while the non-conserved region and loci are in lower probability. Experimental results on BAliBASE3.0 database show that the proposed approach obtains higher fitness value than commonly used alignment algorithm ClustralX does, and has better convergence performance than traditional genetic algorithm.

关 键 词：多重序列比对 协同进化算法 选择算子 交叉算子 变异算子 保守区域 

分 类 号：O242.23[理学—计算数学]