结合蛋白质相互作用数据进行基因表达数据聚类  

Incorporating Protein-protein Interactions Knowledge in Clustering Gene Expression Data

作　　者：黎刚果[1] 王正志[1] 

机构地区：[1]国防科学技术大学机电工程与自动化学院自动化所,长沙410073

出　　处：《生物信息学》2009年第4期280-283,共4页Chinese Journal of Bioinformatics

基　　金：国防科学技术大学创新资助(080320)

摘　　要：提出了一种蛋白质相互作用的相似性度量,将其与基因表达数据的相似性度量相结合,定义了一种融合的距离度量,并且将这种融合的距离度量用于改进现有的K-means聚类方法。经过实际数据的检验,改进后的K-means方法比常用的其它几种聚类方法具有更好的效果,说明结合蛋白质相互作用数据可以使得基因表达聚类的结果更有生物意义。A similarity measure between genes with protein - protein interactions is proposed. On the basis of it, the combined dissimilarity measure is defmed. The combined distance measure is introduced into K - means method and evaluated by a real dataset. The improved K - means method outperforms other clustering methods, which shows that when clustering gene expression data, it can help to get some more bio- logical meaningful results to incorporate protein- protein interactious into the gene expression data.

关 键 词：蛋白质相互作用 基因表达数据 聚类 

分 类 号：Q786[生物学—分子生物学]