复合编码支持向量机预测蛋白质二级结构  

Predicting the secondary structure of protein based on complex coding method and Support Vector Machine

作　　者：郑欣亚[1] 马文丽[1,2] 陈启龙[1,3] 郑文岭[1,2] 

机构地区：[1]上海大学电子生物技术研究中心,上海200072 [2]南方医科大学基因工程研究所,广州510015 [3]黄山学院生命与环境科学学院,黄山245021

出　　处：《微计算机信息》2009年第13期276-278,共3页Control & Automation

基　　金：基金申请人:马文丽;项目名称:基因组研究与基因分离中DNA阵列技术;基金颁发部门:国家自然科学基金委(39880032)

摘　　要：蛋白质结构与功能一直是生命科学的研究重点。尽管蛋白质二级结构的预测已得到广泛的应用,但其预测的精度一直受到算法的制约。在本文中,采用复合编码代替传统的氨基酸编码方式,结合氨基酸疏水性对蛋白质结构的影响,提出一种新的支持向量机算法。使用7倍交叉验证表明,本算法提高了二级蛋白质结构预测的准确性,并节约了计算资源。The structure and function of proteins are sustained largely by different types of studies among life science. Although the prediction of protein secondary structure was application and perform many biological functions that are essential for sustaining life, however, the predicted precision still were inhibited by algorithm. In this paper, a novel Support Vector Machine （SVM） were de- signed for predicting the secondary structure of protein, which use complex coding methods to replace PSSM amino acid coding ways, and integrate the hydrophobic of amino acid into protein structure. The results based on 7 Cross-Validation indicated that this algorithm can improve the accuracy of predictor, and reduce the computing resource.

关 键 词：复合编码 二级结构 三元支持向量机 交叉验证 

分 类 号：Q617.6[生物学—生物物理学]