Development of novel computational methods for simulation of biological systems

AEI 43

Victor M. Anisimov, victor@outerbanks.umaryland.edu, Department of Pharmaceutical Sciences, University of Maryland, 20 Penn Street, Room 601, Baltimore, MD 21201
Classical force fields are an indispensable tool in elucidation of mechanisms of bio-molecular processes. Despite of their utility fixed-charge models have limited applicability for the environments of different polarity. In response to this challenge development of CHARMM Drude polarizable force field has been undertaken. Comprehensive parameter optimization procedure was developed targeting experimental and quantum-mechanical properties of the diverse set of model compounds based on gas-phase data, thermodynamic properties of neat liquids and crystals, and free energies of hydration. Simultaneously the next generation force field is introduced where charge transfer and polarization are described at the QM level. Computational bottleneck characteristic for conventional QM methods is bypassed due to variational finite LMO approximation. Geometry optimization of biological macromolecules beyond hundred thousands of atoms is reported. The complex nature of protein-ligand complexes involving transition metals and highly charged ions, and the problem of charge dependence on conformation of ligands are addressed through all-atom QM molecular dynamics and flexible-ligand QM docking.