Structural Information-Based Method for the Efficient and Reliable Prediction of Oligopeptide Conformations.

Predictions of structures of biomolecules are challenging due to the high dimensionalities of the potential energy surfaces (PESs) involved. Reducing the necessary PES dimensionality is helpful for improving the computational efficiency of all relevant structure prediction methods. For that purpose, a systematic analysis of the backbone dihedral angles (DAs) in the low energy conformations of amino acids, di-, tri-, and tetrapeptides is performed. The analysis reveals that the DAs can be represented by a set of discretized values. Moreover, there are rules limiting the combinations of neighboring DA states. The DA combination rules are used to formulate a path matrix scheme for locating the low energy conformations of peptides. Comparing with the full DA combinations, the PES dimensionality in the path matrix method is reduced by a factor of 2.5n, where n is the number of amino acid residues in a peptide. The path matrix method is validated by applications to find the conformations of representative tri-, tetra-, and pentapeptides and comparison with the best literature results. All the tests show that the path matrix method is very efficient and highly reliable by producing the best search results for the low energy peptide conformations.