A Parallel Algorithm for Mining Maximal Cohesive Subgraphs

Robust and scalable techniques for mining patterns or subgraphs in protein protein interaction (PPI) networks can help identify functionally relevant and coherent subnetworks. Recently, researchers have focused on integrating genes attributes with the protein-protein interaction networks for mining connected subnetworks whose genes are similar in a subset of attributes. However, most of the proposed approaches assume that these subnetworks are dense. While detecting dense and cohesive subnetworks is desirable, the density factor can prevent these algorithms from reporting highly cohesive subgraphs which are not particularly dense. In this paper, we propose a parallel algorithm for mining maximal cohesive subgraphs from node-attributed networks. Experiments on two real interaction networks and gene expression attributes demonstrate the effectiveness of the proposed algorithm. Moreover, biological enrichment analysis of the reported patterns show that the patterns are biologically relevant and enriched with known biological processes and KEGG pathways.