Reconstruction of Dynamic Gene Regulatory Networks for Cell Differentiation by Separation of Time-course Data

Recently, dynamic Bayesian network (DBN) model is widely used for estimating gene regulatory networks (GRNs) from time-course gene expression data. Ordinary DBNs estimate only a single network using the whole timecourse data. However, some GRNs, such as cell differentiation, dynamically change their network structures due to chromatin remodeling. In this papers we present a method to estimate such dynamic GRNs that follow the dynamic changes of the regulations in adipocyte differentiation by separating time-course data. We analyzed the estimated GRNs and confirmed that the GRNs showed the dynamic changes in adipocyte regulation. The result shows that our method can identify the regulatory relationships of the genes that are dynamically changing during adipocyte differentiation by separating the time-course data.