Phase space and collective variable based simulation methods for studies of rare events

摘要

In this work, we review molecular simulation methods for studies of rare events. Specifically highlighted are recent advances and biological applications of methods that utilise unbiased sampling in full phase space to discover reactive trajectories and their ensembles as well as methods that utilise biased sampling along collective variables to efficiently sample free energy landscapes. Among phase space methods, we discuss transition path sampling and its variants and highlight emerging themes in analyses of reactive trajectories via transition path theory, reactive islands, and Lagrangian descriptors. Among collective variable methods, we discuss techniques that allow efficient sampling in a high-dimensional CV space, either via replica-based approaches or by invoking an independent sampling protocol for each CV. Specifically, we highlight two variants of the metadynamics method, bias-exchange and parallel-bias metadynamics, and a hybrid method termed temperature-accelerated sliced sampling, that have been designed to overcome limitations of related approaches.