Model reduction of rigid-body molecular dynamics via generalized multipole potentials

Area of Science:

• Computational chemistry
• Molecular dynamics
• Statistical mechanics

Background:

• Uncertainty quantification in coarse-grained (CG) molecular dynamics presents significant challenges.
• Model reduction strategies are crucial for simulating large-scale molecular systems.

Purpose of the Study:

• Investigate the role of perturbation theory in coarse-grained molecular dynamics model reduction.
• Analyze the accuracy of generalized multipole potentials for rigid body coarse-graining.
• Develop methods to understand and mitigate errors in CG models.

Main Methods:

• Applied perturbation theory to generalized multipole potentials for rigid body systems.
• Developed a hierarchy of CG models from "point molecule" approximations to exact dynamics.
• Derived asymptotic error estimates for approximate molecular potential energies.

Main Results:

• Omitting rotational energy significantly increases error in coarse-grained models.
• Incorporating symmetry considerations improves the accuracy of "point-molecule" approximations.
• The multipole framework systematically connects CG models to atomistic dynamics.

Conclusions:

• Perturbation theory provides a framework for uncertainty quantification in CG molecular dynamics.
• Careful treatment of rotational energy and symmetry is essential for accurate CG models.
• The generalized multipole potential approach offers controllable accuracy in model reduction.