Impact analysis algorithms for biological interaction networks

Abstract

Gene expression profiling (GEP) and genome-wide association studies (GWAS) are powerful tools that can provide list of genes that are related to the pathogenesis of a disease, but it is still a challenge to understand how multiple genes that have modest association with the phenotype interact and contribute to it. For this purpose, it is required to consider molecular profiles with biological interactions. In this work, we proposed two active module identification methods: an active subnetwork search method based on genetic algorithm and a network propagation method. We aimed to understand affected paths in interaction networks and reveal underlying disease mechanisms. We applied our methods to rheumatoid arthritis, intracranial aneurysm and Behçet’s disease GWAS datasets. The proposed methods could successfully identify pathways that are known to be related to the diseases, and extract new mechanisms.