Michael Cusick, Julie Sahalie, David Hill, Mike Boxem, Jean-Francois Rual, Irma Lemmens, Anne-Ruxandra Carvunis, Chenwei Lin, Niels Klitgord, Amelie Dricot, David Szeto, Muhammed Yildirim, Nicolas Simonis, Fana Gebreab, Jan Tavernier, Marc Vidal, Alex Smolyar, Tong Hao, Tomoko Hirozane-Kishikawa, Pascal Braun
[
International Worm Meeting,
2007]
Systematic identification of protein-protein interactions is of immense value for cell and systems biology. Large-scale mapping of protein-protein interactions provides novel functional categorization of uncharacterized proteins and characterized proteins alike, and informs upon the emergent global properties of cellular networks. In pursuit of these aims we will present an expanded version of the C. elegans protein-protein interaction network, or interactome, map. Our last release, worm interactome 5 (WI5), focused on a subset of predicted worm proteins that have a clear ortholog in other multicellular organisms, and contained 1,039 high-confidence interactions representing ~3-5% of the predicted C. elegans interactome. Here we attempted to test all potential interactions in a matrix of ~11,000 x ~11,000 ORFs using an improved highly specific version of the high-throughput yeast two-hybrid (HT-Y2H) system, covering ~25% of the search space of the worm interactome. We identified 1,899 high-quality interactions, of which 1,643 are new. The technical quality of the interactions is being assessed by an orthogonal assay (MAPPIT), while the biological relevance is being estimated by comparing co-expression and Gene Ontology annotations of interacting pairs. The comparison of this new screen with the WI5 HT-Y2H experiment, which was performed with a very different strategy, allows estimations of the coverage and reproducibility of these two screens. Previous interactions from high-throughput, medium-throughput and literature curated datasets have been reprocessed according to new quality criteria and incorporated with the new HT-Y2H screen to generate a consolidated map (WI8) of 4,196 high quality binary interactions between 2,624 protein coding genes. Integration of WI8 with expression, genetic interaction, phenotypic and interolog data is being used to delineate functional modules, which will lead to new functional predictions for numerous genes.