(746h) Efficient Simulation of the Hydrophobic Effect in Biological Self-Assembly

The assembly of proteins into complexes and higher-order structures occurs on the scale of nanometers, where the thermodynamics of solvation are not well described by macroscopic thermodynamics and the microscopic structure of water determines the hydration layer around a protein. In atomistic as well as coarse-grained simulations, solvent is often taken into account explicitly to describe the hydrophobic effect and protein surface interactions, which severely limits the time scales accessible in simulation because of integrating the water degrees of freedom. Inspired by recent work [1], we propose a novel implementation of an implicit solvation scheme in Molecular Dynamics on GPUs, which is based on the theory for the hydrophobic effect of Lum, Chandler and Weeks for the nanometer-scale solvation and a smoothed particle description on larger scales [2]. This model captures the hydrophobic effect due to exclusion of solvent fluctuations from small and large cavities accurately, while being computationally efficient. The goal is to routinely do simulations on millisecond time scales in biological self-assembly on small clusters of GPUs.

 [1] S. Vaikuntanathan, G. Rotskoff, A. Hudson, and P. L. Geissler, â??Necessity of capillary modes in a minimal model of nanoscale hydrophobic solvation,â? Proc. Natl. Acad. Sci., vol. 113, no. 16, pp. E2224, 2016.

[2] H. Lei, C. J. Mundy, G. K. Schenter, and N. K. Voulgarakis, â??Modeling nanoscale hydrodynamics by smoothed dissipative particle dynamics,â? J. Chem. Phys., vol. 142, no. 19, p. 194504, 2015.