Many-Body Coarse-Grained Interactions Using Gaussian Approximation Potentials.

Area of Science:

• Computational Chemistry
• Materials Science
• Statistical Mechanics

Background:

• Coarse-grained (CG) models simplify molecular complexity for large-scale simulations.
• Traditional CG models often rely on pair potentials, limiting accuracy for many-body interactions.
• Accurate free energy calculations are crucial for understanding molecular behavior and predicting material properties.

Purpose of the Study:

• To introduce a novel computational framework for describing general many-body coarse-grained interactions.
• To model the free energy surface of molecular liquids using a cluster expansion.
• To enhance the accuracy of CG models by incorporating higher-order interaction terms.

Main Methods:

• Developed a computational framework for many-body CG interactions.
• Modeled free energy surfaces as cluster expansions (monomer, dimer, trimer terms).
• Inferred free energy contributions from all-atom molecular dynamics (MD) data using Gaussian Approximation Potentials (GAP), a machine learning model based on Gaussian process regression.

Main Results:

• The developed CG model significantly improves accuracy compared to models using only pair potentials.
• The framework successfully captures monomer, dimer, and trimer contributions to the free energy.
• The CG model, while computationally intensive, can be faster than all-atom simulations for specific applications.

Conclusions:

• The new computational framework provides a more accurate description of many-body CG interactions.
• This approach advances the accuracy of free energy calculations in molecular liquids.
• The method offers a promising alternative for efficient and accurate molecular simulations, particularly in biomolecular contexts.