基于OET-KNN算法的蛋白质二级结构类型预测  

Protein secondary structural classes prediction based on OET-KNN model-ing

作　　者：邱望仁[1] 肖绚[1] 林卫中[1] 

机构地区：[1]景德镇陶瓷学院信息工程学院,江西景德镇333403

出　　处：《计算机工程与应用》2008年第29期204-206,210,共4页Computer Engineering and Applications

基　　金：国家自然科学基金No.60661003;江西省自然科学基金No.0611060;江西省青年科学家(井冈之星)培养计划资助~~

摘　　要：蛋白质二级结构类型预测是当今生物信息学研究的热点之一。利用氨基酸数字编码模型将氨基酸序列转换成数字信号,根据LZ复杂度的算法计算了氨基酸的伪氨基酸成分,再对伪氨基酸成分用OET-KNN算法进行分类预测。Jackknife测试结果表明该算法能使得预测成功率有较大的提高。Protein secondary, structure prediction is the hot of bioinformatics.In this paper,a novel method based on optimal evidence-theoretic K nearest neighbor（OET-KNN） algorithm has been introduced,in which,based on encoding the amino acid sequence into digital signals,the pseudo amino acid composition is incorporated with the complexity through the LZ＇s algorithm.The result of these pseudo-amino acids shows that the prediction success rate is improved.

关 键 词：蛋白质 二级结构型预测 K-近邻算法 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]