Use of amino acid environment-dependent substitution tables and conformational propensities in structure prediction from aligned sequences of homologous proteins I. Solvent accessibility classes

Buried and exposed residues are predicted by composing amino acid substitution patterns and mean propensities for the two solvent accessibility classes with the amino acid residues at equivalent sites in aligned sequences of homologous proteins. In a study of 13 protein families, the accuracy of the prediction is around 77% (the correlation coefficient between the predicted and observed accessibility classes is 0·52). The environment-dependent amino acid substitution tables are especially important in prediction of buried hydrophilic and exposed hydrophobic residues, which are not well predicted with propensities alone. Since the prediction is site-specific in the sense that any averaged properties over neighbouring residues are not required, the results can be used for the prediction of secondary structures by detecting periodicity in the sequence of buried and exposed classes.