Detecting Co-Regulatory Modules from Human Regulatory Network by Randomly Walking Between Regulator and Gene Modules

Gene expression is jointly regulated by microRNAs and transcriptional factors. As such, constructing a regulatory network for microRNAs and transcriptional factors and analyzing their combinatorial effects are vital to understand living organisms. Co-regulatory modules, including functional homogeneous microRNAs, transcriptional factors, and genes, provide insights into coordinate regulation. In this paper, we propose a random walk with restart between regulator and gene modules (RWRRGM) method to detect co-regulatory modules from a human regulatory network. The network integrates large, heterogeneous data, including transcriptional regulation, post-transcriptional regulation, and gene-gene interaction. RWRRGM first identifies regulator and gene modules by greedily expanding seed nodes and then walks on the identified modules randomly. Finally, functional homogeneous regulator and gene modules are integrated to form co-regulatory modules. RWRRGM-based modules exhibit higher enrichment in gene ontology terms and known pathways than modules predicted by other methods.