Abstract
Many bioinformatics methods have been proposed for reducing the complexity of large gene or protein networks into relevant subnetworks or modules. Yet, how such methods compare to each other in terms of their ability to identify disease-relevant modules in different types of network remains poorly understood. We launched the ‘Disease Module Identification DREAM Challenge’, an open competition to comprehensively assess module identification methods across diverse protein–protein interaction, signaling, gene co-expression, homology and cancer-gene networks. Predicted network modules were tested for association with complex traits and diseases using a unique collection of 180 genome-wide association studies. Our robust assessment of 75 module identification methods reveals top-performing algorithms, which recover complementary trait-associated modules. We find that most of these modules correspond to core disease-relevant pathways, which often comprise therapeutic targets. This community challenge establishes biologically interpretable benchmarks, tools and guidelines for molecular network analysis to study human disease biology.
Highlights
Complex diseases involve many genes and molecules that interact within cellular networks[1,2,3]
We developed a panel of diverse, human molecular networks for the challenge, including custom versions of two protein–protein interaction and a signaling network extracted from the STRING14, InWeb[15] and OmniPath[16] databases, a co-expression network inferred from 19,019 tissue samples from the Gene Expression Omnibus (GEO) repository[17], a network of genetic dependencies derived from loss-of-function screens in 216 cancer cell lines[18,19] and a homology-based network built from phylogenetic patterns across 138 eukaryotic species[20,21] (Methods)
Solvers were asked to identify a single set of non-overlapping modules by sharing information across the six networks, which allowed us to assess the potential improvement in performance offered by emerging multi-network methods compared to single-network methods
Summary
A community challenge to assess network module identification methods. In Sub-challenge 1, solvers were asked to run module identification on each of the provided networks individually (single-network module identification). Solvers were asked to identify a single set of non-overlapping modules by sharing information across the six networks (multi-network module identification), which allowed us to assess the potential improvement in performance offered by emerging multi-network methods compared to single-network methods. In both sub-challenges, predicted modules had to be non-overlapping and comprise between 3 and 100 genes. Nature Methods | VOL 16 | SEPTEMBER 2019 | 843–852 | www.nature.com/naturemethods
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
