Adaptive Boundaries in Multi-Resolution Simulations

Biomolecular simulations generally require a compromise between forcefield resolution and computational efficiency. Methods that combine multiple levels of resolution promise to extend the ability of simulations to handle bigger systems and more complex processes.We have developed an explicit/continuum solvent model that is able to reproduce the effects of explicit biomolecular solvation while only including a fraction of the molecules that would otherwise be required. We are extending this approach as a general multi-resolution model where both solvent and other, more complex, molecules change representation as they move across the boundaries of the explicit and continuum domains.This model includes: (1) new boundary methods that accurately reproduce thermodynamic and kinetic properties in the explicit region exactly, within statistical error, as if they were taken from large, fully explicit simulations; (2) adaptive boundaries that spatially alter the level of detail in response to an evolving molecular environment; and (3) a grand canonical control of molecular components, relaxing the density of chemical species as the simulation progresses (important for studies in crystallization, assembly, etc.).Our current work is aimed at surmounting a considerable challenge facing all multi-resolution models: transferability to new systems and arbitrary geometries. Overcoming this challenge will be essential in making these multi-resolution models ready to be used ‘out of the box’ for a range of problems in biophysics.