Network-Based Predictions and Simulations by Biological State Space Models: Search for Drug Mode of

• 作 者: Rui Yamaguchi Seiya Imoto Satoru Miyano
• 机 构: Human Genome Center, Institute of Medical Science, University of Tokyo, 4-6-1 Shirokanedai, Minato-k
• 出 处: 计算机科学技术学报（英文版）
• 发表日期: 2010年1期
• 关键词: gene networks state space models time-course gene expression data

Since time-course microarray data are short but contain a large number of genes, most of statistical models should be extended so that they can handle such statistically irregular situations. We introduce biological state space models that are established as suitable computational models for constructing gene networks from microarray gene expression data. This chapter elucidates theory and methodology of our biological state space models together with some representative analyses including discovery of drug mode of action. Through the applications we show the whole strategy of biological state space model analysis involving experimental design of time-course data, model building and analysis of the estimated networks.