非负矩阵分解算法在胃癌基因表达数据分类中的应用

摘要

依据基因表达数据,利用计算机技术对其样本进行分类并找到在肿瘤组织中特异表达的基因,使之能够对疾病的治疗和生物医学研究起到有益的参考和借鉴作用.本文采用非负矩阵分解算法,对基因表达数据进行分析,然后对分解后得到的基矩阵中各集合基因进行比较,找出在肿瘤组织中有特异表达的基因,并做出生物解释.以胃癌基因表达数据为例进行实验,结果表明了该方法的可行性和有效性.%Based on the gene expression data, we correctly finished the classification of samples and find the genes with special expressions in tumor tissues by computer technology, which is critically important for diseases clinical diagnosis and biomedical researches.Recently,the clustering methods have been frequently used in tumor classification and diagnosis.In this paper,we use a two-way clustering approach called Non negative matrix factorization(NMF).Two steps adopted in gene expression data.The first step deal with the gene expression data with this approach.The second step discover the genes have great relationship with the disease by comparing each cols of the matrix obtained after faclorizalion.The two steps are applied to the gastric cancer gene expression data,which shows the feasibility and effectiveness of the method proposed