Invited Talks

MAPPING AND ANALYSIS OF THE HUMAN INTERACTOME NETWORK

Abstract
We have mapped a first version of the human interactome network using a high-throughput Y2H (HT-Y2H) technology. Our data set, CCSB-HI1 is high in specificity and adds ~2700 new protein-protein interactions to existing interactome maps. CCSB-HI1 interactions are enriched for correlations with mRNA coexpression, presence of shared conserved cis regulatory DNA motifs, shared phenotypes and shared function.

A systematic quantitative examination of various existing human interactome maps shows that, contrary to existing notion, high-throughput Y2H maps are in fact higher in specificity than the composite information obtained from curating literature containing a large number of papers describing one or a few interactions at-a-time.

Furthermore, combined experimental and computational modeling of repeat trials of a HT-Y2H screen predicts the size of the Y2H-detectable human interactome and demonstrates the feasibility of mapping a nearly complete set of human interactions through multiple screens in a reasonable time frame.

Novel candidate disease genes and associated hypotheses emerge for more than 300 interactions involving disease proteins from this data set.

This existing interactome map can be used to begin to investigate how cellular networks are perturbed in disease. For example, from analysis of a draft interactome map of Epstein-Barr virus proteins with human proteins that we generated, we find that EBV proteins tend to target highly connected or hub proteins in the human interactome, and moreover, proteins that are central in the network, having relatively short paths to other proteins in the network.