Description (provided by applicant): The protein-protein docking problem is one of the focal points of activity in computational structural biology. The 3D structure of a protein-protein complex, generally, is more difficult to determine experimentally than the structure of an individual protein. Adequate computational techniques to model protein interactions are important because of the growing number of known protein 3D structures, particularly in the context of structural genomics. The project will improve our understanding of fundamental properties of protein interaction and will facilitate development of better tools for prediction of protein complexes. The Specific Aims of the project are: (1) Advanced docking algorithm, (2) Resource databases, and (3) Integrated web-based environment. The long-term goals are: (a) development of an automated tool for a reliable modeling of protein interactions, which will account for dynamic changes in the molecular structures and kinetics of protein association and (b) utilization of this tool to understand principles of protein interaction. The ultimate goal is to recreate the network of protein interactions in genomes and understand the structure-base mechanisms of these interactions. The systematic, detailed description of these interactions will provide insights into the basic principles of life processes at the molecular level. The focus of the proposal is an integrated system of resources for studying protein-protein 3D interactions. An existing docking procedure will be developed further to make it more adequate to the challenges of structural modeling of protein-protein complexes. The development will make use of the rapidly growing body of experimentally determined structures of protein-protein complexes. The procedure will be used to generate docking datasets for the development of modeling capabilities. The core dataset consists of regularly updated and annotated co-crystallized protein-protein structures. The database of experimentally determined and simulated unbound complexes will be further expanded upon the core dataset. It will serve as a comprehensive benchmark set for the development of docking techniques. The database of protein-protein models will provide a unique expansion of the core dataset for development of docking capabilities in protein modeling, including genome-wide studies. The database of docking decoys will provide the community-wide testing ground for new scoring functions.