随机矩阵理论在肝癌基因功能模块识别中的应用

采用随机矩阵理论方法研究了肝癌的基因表达网络。通过标准误差分析，得到了从富含噪声的肝癌基因网络中分离出真实肝癌基因网络的、去躁最充分的关联系数，分析了由此获得的基因表达网络的13个基因功能模块，发现这些模块与肝癌的产生和发展有密切关系。基于随机矩阵理论的方法克服了以往模块识别方法带有主观因素且不能去除噪声因子的缺陷，是一种有效去除随机噪声、识别基因模块、简化基因网络的方法。由于基因数目的众多及细胞生物过程的复杂性，从整体的角度系统研究肝癌基因表达谱，对理解肝癌分子机制和探索新的治疗方法有重要的现实意义。

Application of Random Matrix Theory to Microarray Data for Discovering Functional Gene Modules of Hepatocellular Carcinoma

The function modules of hepatocellular carcinomas (HCC) gene expression network was identified by Random Matrix Theory (RMT). The standard deviation of the eigen-value spacing distribution of the expression correlation matrices to the two RMT distributions was used to identify the transition, where the random components were ultimately removed and the correlation matrix contains the clear and important modular information. By analyzing the lager 13 modules revealed by RMT, It was found that these models were closely related to the form and development of hepatocellular carcinomas. The RMT method, having the advantages of avoiding the objective effects and removing the noise caused by experiments, can be an effective way to identify gene functional modules from the complex gene expression networks. Because of the large number of genes and the complexity of cell biological processes, the systematic study of HCC using RMT from an integral perspective helps to understand the mechanisms of hepatocarcinogenesis at molecule level and to advance effective therapy methods.