Motivation: Considerable attention has been directed in recent years toward the development of methods for the contextual analysis of expression data using interaction networks. Of particular interest has been the identification of active subnetworks by detecting regions enriched with differential expression. In contrast, however, very little effort has been made toward the application of comparable methods to other types of high-throughput data.
Results: Here, we propose a new method based on co-clustering that is specifically designed for the exploratory analysis of large-scale, RNAi-based functional screens. We demonstrate our approach by applying it to a genome-scale dataset aimed at identifying host factors of the human pathogen, hepatitis C virus (HCV). In addition to recovering known cellular modules relevant to HCV infection, the results enabled us to identify new candidates and formulate biological hypotheses regarding possible roles and mechanisms for a number of them. For example, our analysis indicated that HCV, similar to other enveloped viruses, exploits elements within the endosomal pathway in order to acquire a membrane and facilitate assembly and release. This echoed a number of recent studies which showed that the ESCRT-III complex is essential to productive infection.
Contact: gonzalez@bio.ifi.lmu.de
Supplementary information: Supplementary data are available at Bioinformatics online.