Mass-scaling replica-exchange molecular dynamics optimizes computational resources with simpler algorithm

On the use of mass scaling for stable and efficient simulated tempering with molecular dynamics

Simulated tempering (ST) is a generalized-ensemble algorithm that employs trajectories exploring a range of temperatures to effectively sample rugged energy landscapes. When implemented using the molecular dynamics method, ST can require the use of short time steps for ensuring the stability of trajectories at high temperatures. To address this shortcoming, a mass-scaling ST (MSST) method is presented in which the particle mass is scaled in proportion to the temperature. Mass scaling in the MSST method leads to velocity distributions that are independent of temperature and eliminates the need for velocity scaling after the accepted temperature updates that are required in conventional ST simulations. The homogeneity in time scales with changing temperature improves the stability of simulations and allows for the use of longer time steps at high temperatures. As a result, the MSST is found to be more efficient than the standard ST method, particularly for cases in which a large temperature range is employed. © 2016 Wiley Periodicals, Inc.

Low-mass molecular dynamics simulation for configurational sampling enhancement: More evidence and theoretical explanation

It has been reported recently that classical, isothermal-isobaric molecular dynamics (NTP MD) simulations at a time step of 1.00 fs of the standard-mass time (Δt=1.00 fssmt) and a temperature of ≤340 K using uniformly reduced atomic masses by tenfold offers better configurational sampling than standard-mass NTP MD simulations at the same time step. However, it has long been reported that atomic masses can also be increased to improve configurational sampling because higher atomic masses permit the use of a longer time step. It is worth investigating whether standard-mass NTP MD simulations at Δt=2.00 or 3.16 fssmt can offer better or comparable configurational sampling than low-mass NTP MD simulations at Δt=1.00 fssmt. This article reports folding simulations of two β-hairpins showing that the configurational sampling efficiency of NTP MD simulations using atomic masses uniformly reduced by tenfold at Δt=1.00 fssmt is statistically equivalent to and better than those using standard masses at Δt=3.16 and 2.00 fssmt, respectively. The results confirm that, relative to those using standard masses at routine Δt=2.00 fssmt, the low-mass NTP MD simulations at Δt=1.00 fssmt are a simple and generic technique to enhance configurational sampling at temperatures of ≤340 K.

On the use of a weak-coupling thermostat in replica-exchange molecular dynamics simulations

In a molecular dynamics (MD) simulation, various thermostat algorithms, including Langevin dynamics (LD), Nosé-Hoover (NH), and weak-coupling (WC) thermostats, can be used to keep the simulation temperature constant. A canonical ensemble is generated by the use of LD and NH, while the nature of the ensemble produced by WC has not yet been identified. A few years ago, it was shown that when using a WC thermostat with particular values of the temperature coupling time for liquid water at ambient temperature and pressure, the distribution of the potential energy is less wide than the canonical one. This led to an artifact in temperature replica-exchange molecular dynamics (T-REMD) simulations in which the potential energy distributions appear not to be equal to the ones of standard MD simulations. In this paper, we re-investigate this problem. We show that this artifact is probably due to the ensemble generated by WC being incompatible with the T-REMD replica-exchange criterion, which assumes a canonical configurational ensemble. We also show, however, that this artifact can be reduced or even eliminated by particular choices of the temperature coupling time of WC and the replica-exchange time period of T-REMD, i.e., when the temperature coupling time is chosen very close to the MD time step or when the exchange time period is chosen large enough. An attempt to develop a T-REMD replica-exchange criterion which is likely to be more compatible with the WC configurational ensemble is reported. Furthermore, an exchange criterion which is compatible with a microcanonical ensemble is used in total energy REMD simulations.