Comments on "Adaptive Resolution Simulation in Equilibrium and Beyond" by H. Wang and A. Agarwal

Abstract

Wang and Agarwal provide a summary and critical appraisal of the AdResS-scheme for resolution-adaptive molecular simulations. They discuss the basic construction of the scheme, including two variants using interpolations of force and potential, respectively, in the transition zone between regions of coarse and high resolution. These variants are compared, especially in the light of the statistics of the full molecular degrees of freedom which they generate in the subdomain treated with full molecular resolution. They first discuss equilibrium statistics and report on a study showing that AdResS samples the grand canonical ensemble with good accuracy. Then they address the question of how well the scheme can represent long-time non-equilbrium properties, such as conformational changes of a complex molecule embedded in the high-resolution region. They provide arguments why the usual global approach to stochastic thermostating should be abandoned when a systems dynamical properties are of interest and discuss the alternative of a local thermostat that is not active in the high-resolution region, but only provides thermal ajdustment through the interaction of this region with the adjacent low-resolution domain.