Active modules for multilayer weighted gene co-expression networks: a continuous optimization approach

Abstract

MotivationSearching for active connected subgraphs in biological networks has shown important to identifying functional modules. Most existing active modules identification methods need both network structural information and gene activity measures, typically requiring prior knowledge database and high-throughput data. As a pure data-driven gene network, weighted gene co-expression network (WGCN) could be constructed only from expression profile. Searching for modules on WGCN thus has potential values. While traditional clustering based modules detection on WGCN method covers all genes, unavoidable introducing many uninformative ones when annotating modules. We need to find more accurate part of them.\n\nResultsWe propose a fine-grained method to identify active modules on the multi-layer weighted (co-expression gene) network, based on a continuous optimization approach (AMOUNTAIN). The multilayer network are also considered under the unified framework, as a natural extension to single layer network case. The effectiveness is validated on both synthetic data and real-world data. And the software is provided as a user-friendly R package.\n\nAvailabilityAvailable at https://github.com/fairmiracle/AMOUNTAIN\n\nContacts.he@cs.bham.ac.uk\n\nSupplementary informationSupplementary data are available at Bioin-formatics online.