Session: COMP Poster Session:

Over the past decades, various advanced sampling methods have been developed to address the challenge of comprehensive sampling of the configurational space of systems with a rough free energy surface. As adopted by several well-known algorithms, one common idea of accelerating the sampling is to modify the underlying free energy surface. For example, metadynamics adds bias along a user-selected set of collective configurations variable, while expanded ensemble algorithms allow a simulation to vary in Hamiltonian during the simulation, and biasing weight as a function of varying Hamiltonian. While this latter method circumvents the need to define a set of collective variables, the system may get stuck in subspaces of the alchemical variable. To alleviate this issue, in this study we propose to add alchemical variables as an allowed collective variable in metadynamics. In this approach, metadynamics is no longer restricted to only sampling the configurational collective variables, but can sample the alchemical direction as well. This approach may be particularly of use in binding free energy calculations, as alchemical simulations are already being carried out. In this study, we validate our implementation of alchemical metadynamics in PLUMED with test systems/alchemical processes with varying complexities and dimensions of collective variable space, from disappearing a lactic acid molecule from water, to the binding and unbinding event of CB8-G3 host-guest binding complex. One-dimensional alchemical metadynamics correctly recovers the free energy difference between the coupled state and the uncoupled state. In addition, multi-dimensional alchemical metadynamics facilitates configurational sampling during ligand binding calculations by introducing configurational collective variables to circumvent barriers. The method can trivially be combined with other metadynamics-based algorithms implemented in PLUMED.

Michael Shirts

Associate Professor, University of Colorado Boulder

Appearances