Assessment of network module identification across complex diseases.

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.

A scored human protein-protein interaction network to catalyze genomic interpretation.

Genome-scale human protein-protein interaction networks are critical to understanding cell biology and interpreting genomic data, but challenging to produce experimentally. Through data integration and quality control, we provide a scored human protein-protein interaction network (InWeb_InBioMap, or InWeb_IM) with severalfold more interactions (>500,000) and better functional biological relevance than comparable resources. We illustrate that InWeb_InBioMap enables functional interpretation of >4,700 cancer genomes and genes involved in autism.

Atg16L1 T300A variant decreases selective autophagy resulting in altered cytokine signaling and decreased antibacterial defense.

A coding polymorphism (Thr300Ala) in the essential autophagy gene, autophagy related 16-like 1 (ATG16L1), confers increased risk for the development of Crohn disease, although the mechanisms by which single disease-associated polymorphisms contribute to pathogenesis have been difficult to dissect given that environmental factors likely influence disease initiation in these patients. Here we introduce a knock-in mouse model expressing the Atg16L1 T300A variant. Consistent with the human polymorphism, T300A knock-in mice do not develop spontaneous intestinal inflammation, but exhibit morphological defects in Paneth and goblet cells. Selective autophagy is reduced in multiple cell types from T300A knock-in mice compared with WT mice. The T300A polymorphism significantly increases caspase 3- and caspase 7-mediated cleavage of Atg16L1, resulting in lower levels of full-length Atg16Ll T300A protein. Moreover, Atg16L1 T300A is associated with decreased antibacterial autophagy and increased IL-1ß production in primary cells and in vivo. Quantitative proteomics for protein interactors of ATG16L1 identified previously unknown nonoverlapping sets of proteins involved in ATG16L1-dependent antibacterial autophagy or IL-1ß production. These findings demonstrate how the T300A polymorphism leads to cell type- and pathway-specific disruptions of selective autophagy and suggest a mechanism by which this polymorphism contributes to disease.

Genetic and Proteomic Interrogation of Lower Confidence Candidate Genes Reveals Signaling Networks in ß-Catenin-Active Cancers.

Genome-scale expression studies and comprehensive loss-of-function genetic screens have focused almost exclusively on the highest confidence candidate genes. Here, we describe a strategy for characterizing the lower confidence candidates identified by such approaches. We interrogated 177 genes that we classified as essential for the proliferation of cancer cells exhibiting constitutive ß-catenin activity and integrated data for each of the candidates, derived from orthogonal short hairpin RNA (shRNA) knockdown and clustered regularly interspaced short palindromic repeats (CRISPR)-Cas9-mediated gene editing knockout screens, to yield 69 validated genes. We then characterized the relationships between sets of these genes using complementary assays: medium-throughput stable isotope labeling by amino acids in cell culture (SILAC)-based mass spectrometry, yielding 3,639 protein-protein interactions, and a CRISPR-mediated pairwise double knockout screen, yielding 375 combinations exhibiting greater- or lesser-than-additive phenotypic effects indicating genetic interactions. These studies identify previously unreported regulators of ß-catenin, define functional networks required for the survival of ß-catenin-active cancers, and provide an experimental strategy that may be applied to define other signaling networks.

Mu-8: visualizing differences between proteins and their families.

BACKGROUND: A complete understanding of the relationship between the amino acid sequence and resulting protein function remains an open problem in the biophysical sciences. Current approaches often rely on diagnosing functionally relevant mutations by determining whether an amino acid frequently occurs at a specific position within the protein family. However, these methods do not account for the biophysical properties and the 3D structure of the protein. We have developed an interactive visualization technique, Mu-8, that provides researchers with a holistic view of the differences of a selected protein with respect to a family of homologous proteins. Mu-8 helps to identify areas of the protein that exhibit: (1) significantly different bio-chemical characteristics, (2) relative conservation in the family, and (3) proximity to other regions that have suspect behavior in the folded protein. METHODS: Our approach quantifies and communicates the difference between a reference protein and its family based on amino acid indices or principal components of amino acid index classes, while accounting for conservation, proximity amongst residues, and overall 3D structure. RESULTS: We demonstrate Mu-8 in a case study with data provided by the 2013 BioVis contest. When comparing the sequence of a dysfunctional protein to its functional family, Mu-8 reveals several candidate regions that may cause function to break down.