A New Ab Initio Algorithm for Structure Prediction of Peptides, Theory, Software and Database

Abstract We present a new fragment-based algorithm to solve the problem of structural prediction for peptide molecules using ab initio calculation. We developed an operation of multilevel ensemble assembly derived from the theory of statistical physics and adopted the structural optimization at the quantum chemistry level. A novel scheme of ensemble acquisition with higher accuracy and completeness is developed specifically for the basic units that act as the initial elements. The ensemble property of fragments at the junction is used to judge the potential of individuals of each block ensemble for construction. The test results show our algorithm is better than the scheme of structure extraction from the Protein Data Bank, which is widely used in biology. We also establish a mathematical form to describe our algorithm and generalize to a framework of a theory appropriate for structural prediction for general linear molecules. Meanwhile, the trial version of this prediction software and peptide database is published.