Highly parallelizable path sampling with minimal rejections using asynchronous replica exchange and infinite swaps.

Area of Science:

• Computational chemistry
• Molecular dynamics
• Statistical mechanics

Background:

• Capturing rare events in molecular simulations is computationally challenging.
• Path sampling methods accelerate simulations but can still be time-consuming.
• Efficiently obtaining thermodynamic and kinetic data from simulations requires advanced techniques.

Purpose of the Study:

• To develop a highly parallelizable and rapidly converging path sampling protocol.
• To overcome the computational limitations of existing path sampling methods.
• To enable efficient simulation of rare but critical molecular events.

Main Methods:

• Implemented subtrajectory moves with high acceptance rates.
• Utilized asynchronous replica exchange with infinite swaps.
• Developed a path sampling protocol compatible with high-performance computing architectures.

Main Results:

• Demonstrated the protocol's effectiveness on liquid-vapor phase transitions, protein unfolding, and water dissociation.
• Achieved comparable statistical accuracy in days for ab initio simulations, a task previously taking over a year.
• Showcased significant acceleration and rapid convergence compared to standard methods.

Conclusions:

• The new path sampling protocol dramatically accelerates the study of rare molecular events.
• This approach enhances the efficiency of molecular simulations for diverse applications.
• It enables faster and more accurate recovery of thermodynamic and kinetic information.