Semi-Implicit Time Integration with Hessian Eigenvalue Corrections for a Larger Time Step in Molecular Dynamics...

Area of Science:

• Computational chemistry
• Biophysics
• Molecular modeling

Background:

• Molecular dynamics (MD) simulations are crucial for understanding biological processes.
• Explicit time integration methods in MD are limited by small time steps, hindering long-timescale simulations.
• Implicit methods offer larger time steps but suffer from computational overhead and instability.

Purpose of the Study:

• To develop a novel semi-implicit time integration scheme for overdamped Langevin dynamics.
• To address the limitations of existing methods in simulating long-time molecular behaviors.
• To enhance the efficiency of molecular dynamics simulations.

Main Methods:

• Proposed the semi-implicit Hessian correction (SimHec) scheme.
• Focused on Hessian matrices of bonded and nonbonded interactions, cutting off large negative eigenvalues to prevent instability.
• Utilized a narrow band Hessian matrix for efficient parallelized linear solutions.
• Tested the scheme on adenylate kinase interdomain fluctuations and myosin II powerstroke using a coarse-grained protein model.
• Introduced a hybrid SimHec-H method to correct deviations.

Main Results:

• SimHec enabled time steps 50-200 times larger than explicit methods.
• Achieved a speedup factor of 7-30, considering computational overhead.
• SimHec reproduced dynamics similar to explicit methods, with slight acceleration in transition dynamics and reduced bonded potential fluctuations.
• The SimHec-H hybrid method corrected these deviations.

Conclusions:

• The SimHec scheme offers a significant advancement for long-timescale molecular dynamics simulations.
• It provides a robust and efficient alternative to existing time integration methods.
• This method facilitates the study of complex biological processes that were previously computationally prohibitive.