This proposal addresses one of the most important problems in protein chemistry, that of predicting binding sites on the surfaces of biologically significant proteins. The long-term goal of this work is to use the ability to identify binding sites from structural information alone to design new pharmaceutical molecules with improved properties. Specifically, the objectives of this project are to develop statistical analyses of the environments of amino acids on protein surfaces, to identify and quantitate surface abnormalities, and to correlate these with experimentally determined binding sites. If successful, this will represent the first comprehensive method to predict binding sites using only the three-dimensional structure of a protein. Follow-on research will use these methods and approaches in collaborations with established pharmaceutical companies to design new protein and peptide drugs with improved specificities, stabilities, and immunological properties.
