Abstract
Motivation: The circadian regulatory network is one of the main topics of the plant investigations. The intracellular interactions among genes in response to the environmental stimuli of light are related to the foundation of the functional genomics in plant. But to determine what genes and how much do they influence each other in whatever ways of transcriptional binding or physical interaction is not easy by the traditionally biological assays. However, under the synchronous measures of gene expression via microarray chip, in which an amount of the dynamic information is hidden and remains to be discovered, we can detect the interactive relationships of genes in the circadian regulatory network via a systematically dynamic modeling analysis in silico. In this study, a first-order dynamic model called ARX (AutoRegressive with eXternal input) is developed to describe the regulatory mechanism of a target gene from the upstream causality point of view. Based on expression profile and dynamic model of a target gene with the prior biological knowledge, we would deduce the upstream regulatory genes with their regulatory abilities under proper activation delays, and then link up a regulatory pathway. After the successful construction of the circadian regulatory network, we would analyze the system under different assumed biological conditions by performing dry experiments and try to discover the essential characteristics of the circadian regulatory network. In words, we can analyze a genetic regulatory pathway of circadian network from the system biology point of view. Results: Our algorithm focuses on the inference of regulatory abilities of the relevant upstream genes in circadian network, and how do they regulate the downstream genes. Then the system sensitivity analysis is investigated to gain much insight to the circadian regulatory network in Arabidopsis thaliana via microarray data and evaluate the network dynamic model in three respects of Input, Trans, and Cis level from system point of view.