Gene relevance based on multiple evidences in complex networks

• DOI: 10.1093/bioinformatics/btz652
• Publication Date: 2019-08-20T19:16:59Z
• Authors: Noemi Di Nanni; Matteo Gnocchi; Marco Moscatelli; Luciano Milanesi; Ettore Mosca
• Keywords: Relevance; Leverage (statistics); Gene regulatory network; R package

Multi-omics approaches offer the opportunity to reconstruct a more complete picture of molecular events associated with human diseases, but pose challenges in data analysis. Network-based methods for analysis multi-omics leverage complex web macromolecular interactions occurring within cells extract significant patterns alterations. Existing network-based typically address specific combinations omics and are limited terms number layers that can be jointly analysed. In this study, we investigate application network diffusion quantify gene relevance on basis multiple evidences (layers).We introduce score (mND) quantifies biological process taking into account proximity its first neighbours other altered genes. We show mND has better performance over existing finding genes one or layers. also report good performances recovering known cancer The pipeline described article is broadly applicable, because it handle different types inputs: addition datasets, datasets stratified many classes (e.g., cell clusters emerging from single analyses) combination two scenarios.The R package 'mND' available at URL: https://www.itb.cnr.it/mnd.Supplementary Bioinformatics online.