Replica exchange molecular dynamic simulations have been widely used to accelerate sampling in proteins. Several replicas with modified thermodynamic parameters are run in parallel with exchanges, allowing sampling at one thermodynamic state to be shared with simulations performed at other states. We use OpenMM, a MD simulation toolkit which supports GPU acceleration, to run Hamiltonian (HREM) and temperature (TREM) replica exchange methods for two important protein applications: (1) Identifying putative binding locations in protein-ligand systems. We use a flat-bottom potential to restrain the ligand in a specified area and let the ligand sample freely using modified HREM to escape local minima.  Binding sites are accurately identified 70X faster than similar codes on standard processors. (2) Developing predictive relationships between structural and molecular properties of proteins and their propensity to unfold when absorbed on chromatography media. We use TREM and coarse-grained representations of proteins to model local unfolding and larger scale unfolding respectively.

Applications of GPU-accelerated replica exchange molecular dynamic simulations of proteins

244th National Meeting (2012)

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.