【作者】 张世华； 吴凌云； 金光旭； 章祥荪；

【Author】 Shi-Hua Zhang Ling-Yun Wu Guangxu Jin Xiang-Sun Zhang Academy of Mathematics and Systems Science, CAS, Beijing 100080

【机构】 中国科学院数学与系统科学研究院；

【摘要】 本文尝试通过应用一种简单的打分系统和类似于序列比对的动态规划策略来发展一种简单,快速的蛋白质结构分类方法。打分系统是基于一套简单的蛋白结构参数。随后使用动态规划策略在点阵上来搜索相似性轨迹以及根据相似性轨迹的得分计算两个蛋白质结构的相似度。基于每对蛋白质结构的相似度, 用层次聚类算法可以将蛋白质分类。最后通过随机选取属于α,β,α／β,α+β类的五个不同家族的蛋白质结构来检验该方法的有效性。更多还原

【Abstract】 The objective of this study is to develop a novel and fast method for structural classification of proteins using a simple scoring system and dynamic programming which bear an analogy to sequence alignment. The scoring system is based on a simple protein profile parameter which has been used for structural similarity search in somewhere. Then a dynamic programming strategy is used for finding the similarity trace on the dot-matrix and the final degree of similarity of two proteins is computed according to the score of retrieving the similarity trace. Based on the degree of similarity of each pair of proteins, the proteins can be classified by cluster analysis with a hierarchical clustering algorithm. The method is tested over a data set belonging to five different families randomly selected from α, β, α/β, α + β classes.更多还原