A Toolkit to Fit Nonbonded Parameters from and for Condensed Phase Simulations

Area of Science:

• Computational chemistry
• Molecular modeling
• Physical chemistry

Background:

• Atomistic simulations rely on accurate force fields, particularly nonbonded interactions (electrostatic and Lennard-Jones terms).
• Permanent atomic multipoles (MTPs) offer an extension to traditional point-charge (PC) representations.
• Parameterizing force fields involves fitting to ab initio calculations and experimental observables via molecular dynamics (MD) simulations.

Purpose of the Study:

• To introduce a comprehensive computing environment, the Fitting Wizard (FW), for optimizing nonbonded interactions in atomistic force fields.
• To evaluate the performance of both PC-based and MTP-based force fields.
• To assess the accuracy of hydration free energy predictions.

Main Methods:

• Development of the Fitting Wizard (FW) supporting both PC and MTP force field optimization.
• Fitting Lennard-Jones parameters using thermodynamic observables (density, heat of vaporization, hydration free energy) via MD simulations.
• Validation against experimental data for a diverse set of 20 molecules.

Main Results:

• The Fitting Wizard (FW) was successfully implemented for optimizing nonbonded interactions.
• A broad study across 20 molecules achieved a root mean squared deviation of 0.36 kcal/mol for hydration free energies.
• Point-charge (PC) force fields demonstrated comparable accuracy to multipolar (MTP) parametrizations for specific molecular systems.

Conclusions:

• The Fitting Wizard (FW) provides a robust tool for developing accurate atomistic force fields.
• The choice between PC and MTP force fields depends on the specific molecular system and desired accuracy.
• Accurate nonbonded interaction parametrization is crucial for reliable condensed-phase simulations.