We have developed a new reduced model potential function for determination of tertiary protein structure assuming knowledge of secondary structure. The model potential uses both alpha-carbon and beta-carbon distance information and is based upon a synthesis of data base statistics, molecular mechanics potentials, and direct optimization of parameters. A key aspect of development of the potential has been minimization of the native structure using an analytical gradient methodology. To perform global optimization, we have implemented a modified version of the branch and bound algorithm proposed by Floudas and coworkers, which appears to be quite powerful for dealing with this type of model. Promising results have been obtained for a significant number of proteins ranging in size from 40 to 180 amino acid residues.

Back to Workshop Schedule

1996-1997 Mathematics in High Performance Computing