Dynamic force matching: A method for constructing dynamical coarse-grained models with realistic time dependence.

Area of Science:

• Chemical Physics
• Computational Chemistry
• Molecular Dynamics

Background:

• Coarse-grained (CG) models simplify molecular systems but often fail to accurately reproduce dynamics.
• Existing methods for dynamic CG models typically use Markovian dynamics (e.g., Langevin, Brownian), which may not suit all systems.
• The Multi-Scale Coarse-graining (MS-CG) method excels at structural properties but not necessarily dynamics.

Purpose of the Study:

• To develop a novel method for constructing accurate dynamic CG models.
• To address the limitations of Markovian dynamics in CG simulations.
• To create CG models that capture non-Markovian dynamics consistent with all-atom simulations.

Main Methods:

• Proposed a method to convert MS-CG models into dynamic CG models by incorporating fictitious particles.
• These fictitious particles interact with CG degrees of freedom and are subject to Langevin forces.
• Model dynamics are based on nonlinear systems interacting with special heat baths (Zwanzig, 1973).
• Fictitious particle properties are derived from all-atom simulation dynamics analysis.

Main Results:

• The developed dynamic CG models exhibit non-Markovian or near-Markovian behavior.
• The dynamics generated are consistent with the all-atom system used for model construction.
• Tests on simple examples demonstrate the method's ability to produce realistic dynamical CG models.
• Computational requirements for construction and simulation are comparable to standard MS-CG models.

Conclusions:

• The proposed method effectively generates realistic dynamic CG models from MS-CG potentials.
• These models offer a viable approach for simulating the dynamics of very large molecular systems.
• The method provides a pathway to capture essential non-Markovian dynamics in coarse-grained simulations.