Reconstructing atomistic detail for coarse-grained models with resolution exchange

Coarse-grained (CG) modeling has emerged as a promising tool to bridge the gap between the temporal and spatial scales of all-atom (AA) simulations and those of many important biological processes. Resolution exchange, a variant of the replica exchange method, combines the efficiency of CG simulation and the accuracy of AA simulation by swapping configurations between AA and CG simulations. The crucial step in a resolution exchange move is to rigorously reconstruct the high-resolution system from models at coarser resolutions. In this paper, configurational-bias Monte Carlo is adopted as a general method to rebuild the missing degrees of freedom rigorously for CG models and for the first time combined with resolution exchange. The new approach is demonstrated on an alkane and a peptide system. It is found that the efficiency of resolution exchange depends significantly on the quality of the CG model. (c) 2008 American Institute of Physics. [DOI: 10.1063/1.2976663]