Multiscale molecular dynamics and the reverse mapping problem

Multiscale techniques are becoming increasingly important for molecular simulation as a result of interest in increasingly complex problems involving events occurring over multiple time and length scales. Here, inspired by the success of the multiscale quantum mechanics/molecular mechanics (QM/MM) methods, we introduce a hybrid, adaptive resolution, multiscale molecular dynamics method that combines accurate, atomistic, modeling of key regions of the system with a coarse-grained modeling of the remainder of the system. Hybrid multiscale methods must solve the interfacial hand-shaking problem of coupling together different levels of description in different spatial regions of the system; in addition, to implement an adaptive resolution algorithm to correctly model diffusive systems, one must have a procedure in place to dynamically change the representation of a molecule, either from a finer to a coarser level or vice versa. We propose a solution to these problems through a detailed energy analysis and the use of a rotational dynamics to align molecular fragments. The algorithms we propose significantly advance the state-of-the-art and should serve to spur significant advances in our ability to model complex chemical systems.