Gene Regulation Network Modeling via Improved Multi-agent System and Dynamic Bayesian Network

Gene regulation network is playing a more and more important role in system biology, bioinformatics and pharmacology in the era of post-genomic. This paper presents a high efficient and accurate method to model gene regulation network. In the method, we first use an improved Multi-Agent System (Im-MAS) to fuse multiple data sources to generate an initial network, and then we use dynamic Bayesian network learning to generate the final network. Evaluation of the approach using real data sets, including 25 genes´ expression data and transcription factor (TF) binding data, shows significantly improved performance over the method proposed in the previous papers.