While the human genome sequence is known, medical advances will require knowledge of how individual genes and proteins assemble into functional biological networks. These networks are built in part through physical interactions between proteins. Protein interaction networks can now be probed using highthroughput experimental methods, yet challenges remain: data can be noisy and incomplete, and methods developed for simpler model organisms can be difficult to scale up to human. This proposal aims to determine the feasibility of mapping human networks relevant to disease by developing a joint computational / experimental approach to meeting these challenges. Specific aims include developing a statistical metric for confidence in proteomic data from high-throughput two-hybrid screens; developing computational methods for mapping networks cross-species; and generating experimental data with network exploration guided by the computational predictions. If this feasibility study is successful, work in Phase II will focus on building proof-of-principle networks for two specific disease areas. Exemplary areas are cancer, to suggest potential targets for small-molecule and antibody drugs, and infectious disease, to identify host-pathogen interaction. Success in Phase I of this project will result in improved methods for mapping disease-relevant human biological networks. Success in Phase II will result in targets for therapeutic intervention
