DOCKING IN LARGE DATABASES OF MODELED PROTEINS

Medical University Of South Carolina, Charleston SC

Investigators

Linked publications & trials

Abstract

The major goal of the project is to design structure-based procedures for building a network of connections between proteins in genomes. The number of protein-protein interactions in a genome is significantly larger than the number of individual proteins. Moreover, a large part of protein structures will be models of limited accuracy. Thus, the structure-based methods for building this network have to be (a) fast, and (b) insensitive to significant inaccuracies of modeled structures. The precision of these methods may be correlated with the precision of the protein structures lower for less accurate models and higher for more exact models. The networks of connections between proteins have to be built by a combination of experimental studies, knowledge-based data, sequence analysis, and structural approaches. The subject of this proposal is the methodology development for the structure-based studies. The long-term goals are to understand the fundamental principles of protein interaction and to create a structure-based description of entire genomes, with the focus on structure-based modeling of protein pathways, accurate description of dynamics and kinetics of protein interactions, engineering of protein structures with special properties, and computer-aided drug design on modeled structures. The specific aims are: (1) docking of modeled protein structures, (2) binding site prediction and comparison, (3) prediction of interacting and noninteracting proteins, and (4) development of the public resource for modeling protein interactions. The database of protein-protein complexes, that includes models of different accuracy, and the database of protein docking decoys will be used to develop docking and binding site prediction approaches and the accuracy improvement procedures. The intermolecular energy landscape characteristics will be used to identify interacting proteins. The datasets and procedures will become part of the public resource for modeling protein interactions.