MONGKIE: Integrated network analysis and visualization platform for multi-omics data

Abstract

Network-based integrative analysis is a powerful technique for extracting biological insights from multi-layered omics data such as somatic mutations, copy number variations, and gene expression data. However, integrated analysis of multi-omics data is quite complicated and can hardly be done in an automated way. Thus, a powerful interactive visual mining tool supporting diverse analysis algorithms for identification of driver genes and regulatory modules is much needed. Here, we present a software platform that integrates network visualization with omics data analysis tools seamlessly. The visualization unit supports various options for displaying multi-omics data as well as unique network models for describing sophisticated biological networks such as complex biomolecular reactions. In addition, we implemented diverse in-house algorithms for network analysis including network clustering and over-representation analysis. Novel functions include facile definition and optimized visualization of subgroups, comparison of a series of data sets in an identical network by data-to-visual mapping and subsequent overlaying function, and management of custom interaction networks. Utility of MONGKIE for network-based visual data mining of multi-omics data was demonstrated by analysis of the TCGA glioblastoma data. The result of case study showed that network-based analysis in MONGKIE can identify and extract biologically relevant genes and pathways that play important roles in GBM tumorigenesis. MONGKIE is a java-based open-source application built on top of plug-in architecture, thus being platform-independent and easily extensible with additional functionalities. It is available under GNU AGPL v3 license at http://yjjang.github.io/mongkie.