最小生成树用于基因表示数据的聚类算法  

An Algorithm for Clustering Gene Expression Data Using Minimum Spanning Trees

作　　者：杨国慧[1] 周春光[1] 黄艳新[1] 吕慧英[1] 

机构地区：[1]吉林大学计算机科学与技术学院

出　　处：《计算机研究与发展》2003年第10期1431-1435,共5页Journal of Computer Research and Development

基　　金：国家自然科学基金项目 ( 60 175 0 2 4);教育部"符号计算与知识工程"重点实验室资助项目

摘　　要：在生物学研究中 ,需要对植物和动物分类 ,对基因进行分类 ,以获得对种群固有结构的认识 使用聚类分析方法 ,有效地鉴别基因表示数据的模式 ,将它们分组成为由类似对象组成的多个类 ,对研究基因的结构、功能以及不同种类基因之间的关系都具有重要意义 将图论的最小生成树理论引入分子生物学中基因表示数据的聚类分析方法 ,设计了生成树的表示和基于最小生成树的聚类算法 ,证明了该方法对于一些准则函数能够产生全局最优簇 。Genes are divided up for computing and understanding the categories of animals and plants and for getting the knowledge about their connatural structures in the research of the biology It is important that using a clustering analysis method, modes of gene expression data are effectively recognized and they are classified as clusters which are formed by similar objects for studying their structure and function, and the relationship between different species of genes Minimum spanning trees, a graph theoretic approach, is used in clustering gene expression data of molecular biology Expression with spanning trees and clustering analysis method based on minimum spanning trees are designed It is proved that optimum clusters can be obtained by using some rule functions Discussion and evaluation are shown for the algorithm, according to the results of the experiments

关 键 词：基因 聚类 簇 最小生成树 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]