Multiscale implementation of infinite-swap replica exchange molecular dynamics

Area of Science:

• Computational chemistry and biophysics
• Molecular dynamics simulations
• Statistical mechanics

Background:

• Replica exchange molecular dynamics (REMD) is a key technique for enhancing conformational sampling in molecular simulations.
• Traditional REMD relies on Metropolis-Hastings criteria for swap acceptance, which can be inefficient.
• Accelerating sampling is crucial for understanding complex molecular systems.

Purpose of the Study:

• To develop a more efficient and rejection-free implementation of REMD.
• To combine REMD with the heterogeneous multiscale method for optimized sampling.
• To explore the conformational landscape of biomolecules with improved simulation techniques.

Main Methods:

• Implementation of a continuous-time Markov jump process for replica swaps using Gillespie's stochastic simulation algorithm (SSA).
• Integration of REMD-SSA with the heterogeneous multiscale method to achieve the infinite-swap limit.
• Application to alanine dipeptide in vacuum and the C-terminal β-hairpin of protein G in explicit solvent.

Main Results:

• The proposed REMD-SSA method is rejection-free and samples the correct joint distribution.
• Combining REMD-SSA with the heterogeneous multiscale method significantly accelerates swap rates.
• Analysis of the C-terminal β-hairpin of protein G reveals a triple funnel energy landscape with two folded and one misfolded structure stabilized by hydrogen bonds.

Conclusions:

• The new REMD-SSA approach offers an efficient and accurate alternative for molecular dynamics simulations.
• This method is easily implemented and parallelizable, making it broadly applicable.
• The findings provide new insights into the folding pathways and conformational dynamics of protein G.