PREDICTING SOLVENT ACCESSIBILITY - HIGHER ACCURACY USING BAYESIAN STATISTICS AND OPTIMIZED RESIDUE SUBSTITUTION CLASSES

We introduce a novel Bayesian probabilistic method for predicting the solvent accessibilities of amino acid residues in globular proteins, U sing single sequence data, this method achieves prediction accuracies higher than previously published methods. Substantially improved predi ctions-comparable to the highest accuracies reported in the literature to date-are obtained by representing alignments of the example protei ns and their homologs as strings of residue substitution classes, depe nding on the side chain types observed at each alignment position. The se results demonstrate the applicability of this relatively simple Bay esian approach to structure prediction and illustrate the utility of t he classification methodology previously developed to extract informat ion from aligned sets of structurally related proteins. (C) 1996 Wiley -Liss, Inc.