Towards Finding a Better Reaction Coordinate: NtrC and DIMS

Abstract

All atom molecular dynamics is appealing due to the detailed insights towards protein function. But, all-atom molecular dynamics is limited due to the small amount of time spent in the transitions between stable states. Dynamic importance sampling (DIMS) enhances this sampling problem by biasing, with correction, to encourage sampling away from the stable states and to thus improve estimates of barrier crossing times. Nevertheless, DIMS has limitations based on the choice of order parameters that define intermediate states. To improve the representation, in DIMS, of order parameters, we have performed simulations of NtrC and have analyzed the transition populations. Our analysis, using techniques of nonlinear time series analysis and nonlinear dimensionality reduction, has enabled us to determine a new numerical method for finding reduced parameter descriptions of transitions in complex biomolecular systems. The results should enable an iterative approach to finding the best order parameter and at the same time improve convergence for estimates of reaction pathways, kinetics and relative free energy differences.