基于最大信息系数的永久性房颤差异表达基因识别  

Maximal Information Coefficient on Identifying Differentially Expressed Genes of Permanent Atrial Fibrillation

作　　者：刘汉明[1,2] 饶妮妮[1] 李益[1] 罗恒荣[1] 杨阳[1] 杨锋[1] 

机构地区：[1]电子科技大学生命科学学院,成都610054 [2]赣南师范学院计算机科学学院,江西赣州341000

出　　处：《中国生物医学工程学报》2015年第1期8-16,共9页Chinese Journal of Biomedical Engineering

基　　金：国家自然科学基金(81171411)

摘　　要：心房颤动是一种常见的、与年龄相关的心律失常,由其所导致的脑卒中具有高致残率和致死率。通过对高通量基因表达谱的分析,可以帮助理解心房颤动的生物学过程和功能紊乱机制,并发现相关致病基因。新型的非参数统计方法——最大信息系数,在探索双变量之间的关联方面具有独特的优点。利用该方法,发现差异与非差异基因表达值与样本表型之间的关联程度不同,构建差异表达基因识别方法。对永久性房颤基因表达谱GSE2240的分析,识别出41个差异表达基因,其中有14基因是已有工作未发现的新差异表达基因。信号通路和富集分析表明,这些差异表达基因与房颤高度相关。同时,对乳腺癌基因表达数据GSE24037的分析,进一步说明该方法在差异表达基因识别方面的有效性。最大信息系数的非参数特性与抗噪能力,使它非常适合于差异表达基因的识别。The atrial fibrillation is a common arrhythmia disease. It usually causes stroke resulting in high risk on morbidity and mortality. It can be insight into the biological processes, the functional disorders of atrial fibrillation and the genes associated with disease risks by analyzing a microarray data. A proposed novel statistical method named as maximal information coefficient （MIC） has excellent performance in exploring the relationship between two-variables. Based on the degree of relationship between the phenotypes and differentially or undifferentially expressed genes, the statistical method was introduced into the analysis of a microarray of permanent atrial fibrillations （GSE2240）. Total of 41 genes were identified by the method, in which 14 genes are new differentially expressed genes. The differentially expressed genes identified by MIC were demonstrated that they were related with atrial fibrillation diseases, via the analyses of signaling pathway and enrichment in DAVID. Furthermore, MIC was used to analyze a microarray of breast cancer （GSE24037） to test the method. MIC is a non-parametric statistical method with anti-noise, which leads it to be ideally suit for identifying differentially expressed genes.

关 键 词：最大信息系数 心房颤动 差异表达基因 

分 类 号：Q33[生物学—遗传学]