System Identification and Nonlinear Factor Analysis for Discovery and Visualization of Dynamic Gene Regulatory Pathways

DNA microarray time-series provide the information vital to estimate the dynamic regulatory pathways and therefore predict the dynamic interaction among genes in time. While dynamic system identification theory has been applied to many fields of study, due to some practical limitations, this theory has been widely used to analyze DNA microarray time series. In this paper, we describe some of these limitations and propose a hierarchical model utilizing nonlinear factor analysis methods to analyze time-series DNA microarray data and identify the dynamic regulatory pathways. The proposed model is applied to model the eukaryotic cell cycle process using a popular dataset of cell cycle time-series. The results indicate that the proposed method can successfully predict the dynamic pathway involved in the process.