Biological networks integration based on dense module identification for gene prioritization from microarray data

Abstract

BackgroundSelection of genes associated with disease plays an important role in understanding the disease pathogenesis and discovering the therapeutic targets. Network based approaches have been widely used for gene selection or functional module detection. Although microarray technology promotes the study of global changes in the gene expression, it is still not ready for clinical research because of lack of meticulous standards for data collection, analysis and validation. Integrating multiple omics data type can compensate for missing and unreliable information of any single data type and thus increases the reliability of newly discovered knowledge. ObjectiveThe work aims to design a novel dense subgraph based framework to select informative genes from the given gene expression data. The goal here is achieved by integrating the multiple types of biological networks and extracting densely connected components. The work focuses on identifying statistically significant and biologically enriched genes given a gene expression data. Methods and resultThe proposed framework identifies the modules of co-expressed genes and further integrates it with protein-protein interaction (PPI) network to identify modules of strongly interacting genes. The dense gene modules (DGM) embedded in the integrated co-expressed gene network are extracted representing tightly co-expressed gene modules. The proposed approach has been compared with the existing clustering based integration of expression data with PPI network. Besides achieving high prediction accuracy, our proposed approach underlines the robustness and consistency of identifying DGM and selects putative and functionally coherent genes. DiscussionThe co-expression based integration infers interesting biological information associated with the disease and the hub genes that are involved in many biological processes. Such a study reveals that integration of different biological networks for gene selection may provide greater insight in disease related biological process.