Abstract
MicroRNAs are a major class of regulatory molecules involved in a broad range of biological processes and complex diseases. A useful step for understanding their functional role is detecting their influence on genome-wide expression profiles. In this work we use elastic-net regression model that incorporates direct and indirect effects of miRNAs on protein networks for identifying regulatory effect of miRNAs on a list of genes from expression experiment (disease gene signature). Elastic-net regression is used to identify miRNAs whose targets are enriched in disease gene signature. Integrating direct and indirect effects of miRNAs on protein network revealed more significant miRNA enrichment in prostate gene signatures compared to using direct effects alone. Integrating protein networks into regression model revealed significant enrichment of miR-16-1 in upregulated genes in prostate cancer which indicates its putative tumor suppression activity.