关键词: 基因网络, 社区发现, 元数据

Abstract: The identification of disease genes is an important high dimensional statistical problem. The network structure of genes has inspired researchers to shift their attention from single gene identification to genetic module identification. Detecting the genetic module from genetic network is the so-called community detection (or node clustering). Most research in this area only use the topology structure, but neglect the metadata on the nodes. Newman and Clauset proposed a statistical inference based community detection method (NC method) to combine the metadata and topology structure-. In this paper, we take NC method as an example to demonstrate the applications and achievements of statistical methods in genetic network, and discuss the potential improvements from statistical point of view. The analysis of NC method indicates that the statistical thinking and principle play an important role in the analysis of unstructured data, such as the genetic network. The statistical methods need adjustment and optimization according to the characteristics of data and questions of interest.

Key words: Genetic network, Community detection, Metadata