The proposed research focuses on developing efficient Monte Carlo algorithms for sampling conformation of all-atom protein models. Such methods would be beneficial for drug design and predicting ligand binding constraints, solving X-ray and NMR structures of proteins, and in predicting the three-dimensional structure of proteins from their amino acid sequence. A flexible new algorithm, hierarchical Monte Carlo, will be investigated; the specific advantages of this method are that it tolerates large changes in the dihedral angle and accelerates energy evaluations, while still converging to a true Boltzmann distribution. The method will be tested and refined on Met-enkephalin and crambin by making ever large perturbations from the global energy minimum structure and tracking the method's efficiency in recovering the global minimum. The method will then be applied to large proteins such as ribonuclease, and to homology modelling, specifically of the serine proteases.
