APPLICATIONS OF WEIGHTING AND CHIRALITY STRATEGIES FOR DISTANCE GEOMETRY ALGORITHMS TO AN ENTEROTOXIN PEPTIDE ANALOG

Significant improvements are made to a recent algorithm that finds mol ecular conformations using distance geometry on nuclear magnetic reson ance data. Weighting factors for the nearest approximation of the dist ance matrix to a data matrix arc allowed to vary between iterations of the algorithm. These changes are proportional to the error of the dis tance between atoms in the configuration and the nuclear magnetic reso nance data bounds. The weight changes increase the rate of convergence by an order of magnitude. Penalty functions are proposed to ensure th e correct chirality. Numerical results for these modifications and sub sequent energy calculations using CHARMm are given for an analog of th e heat stable (ST) enterotoxin peptide STh produced by E. coli in huma ns.