The study of reaction mechanisms in complex environments, like an enzyme, requires to describe the active site at a computationally expensive quantum mechanical level to allow reaction to happen, but also the inclusion of the protein/solvent matrix explicitly. Hybrid quantum-classical (QM/MM) schemes are the method of choice for these type of studies. However, even for them determination of accurate free energy profiles is not easy due to required extensive sampling of the process and the use of well chosen biased sampling schemes.<br/>In this work we present and thoroughly test a hybrid differential relaxation algorithm (<b>HyDRA</b>), which allows faster equilibration of the classical environment during the nonequilibrium steering of a (bio)chemical reaction. This implementation relies on the differential requirement of relaxation time of each subsystems, treated at the quantum and the classical level respectively.<br/>The results, using an implementation of HyDRA in the AMBER code, show that it allows obtaining accurate free energy profiles with smaller number of independent trajectories and/or faster pulling speeds, thus reducing overall computational cost. We also present a discussion of the underlying reasons for the observed improvement and how it relates to the properties of the reactions under study in the context of Jarzynski's relationship.

Improving efficiency in SMD simulations through a hybrid differential relaxation algorithm

249th National Meeting (2015)

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