Comparison of statistical methods for subnetwork detection in the integration of gene expression and protein interaction network

• DOI: 10.1186/s12859-017-1567-2
• Publication Date: 2017-03-03T00:48:59Z
• Authors: Hao He; Dongdong Lin; Jigang Zhang; Yu-ping Wang; Hong-wen Deng
• Keywords: Subnetwork

With the advancement of high-throughput technologies and enrichment popular public databases, more research focuses bioinformatics have been on computational integration network gene expression profiles for extracting context-dependent active subnetworks. Many methods subnetwork searching developed. Scoring algorithms present a range considerations implementations. The primary goal study is to comprehensively evaluate performance different detection methods. Eleven were selected comprehensive comparison.First, taking into account dependence genes given protein-protein interaction (PPI) network, we simulated microarray data under case control conditions. Then each method was applied identification. Second, large set prostate cancer used assess practical method. Using both simulation studies real application, evaluated in terms recall precision.jActiveModules, PinnacleZ WMAXC performed well identifying with relative high precision recall. BioNet very only precision. As none outperformed other overall, users should choose an appropriate based purposes their studies.