Development and Benchmarking of Open Force Field v1.0.0-the Parsley Small-Molecule Force Field.

Area of Science:

• Computational Chemistry
• Molecular Modeling
• Drug Discovery

Background:

• Classical molecular simulations require accurate force fields for reliable predictions.
• Existing force fields often rely on traditional atom typing, which can limit specificity and increase complexity.
• Developing general-purpose force fields applicable to diverse small molecules is an ongoing challenge.

Purpose of the Study:

• To present a novel methodology for defining and optimizing general force fields for classical molecular simulations.
• To derive and validate the Open Force Field 1.0.0 small-molecule force field, codenamed Parsley.
• To demonstrate the efficiency and accuracy of the developed methodology and force field.

Main Methods:

• Utilized the SMIRKS-native Open Force Field (SMIRNOFF) parameter assignment formalism, moving beyond traditional atom typing.
• Employed the ForceBalance tool for parameter optimization, using reference quantum chemical data including torsion potentials, structures, and vibrational frequencies.
• Leveraged the QCArchive software ecosystem for computation and management of quantum reference data.

Main Results:

• Successfully derived the Parsley (Open Force Field 1.0.0) small-molecule force field.
• Demonstrated improved optimized geometries and conformational energetics compared to training set data.
• Achieved accuracy for liquid properties and binding free energies comparable to existing general force fields.
• Validated performance on 199 protein-ligand systems for relative binding free energy calculations.

Conclusions:

• The presented methodology enables the rapid optimization of novel force fields with minimal human intervention.
• The Parsley force field shows competitive accuracy for molecular simulations, including binding free energy predictions.
• The SMIRNOFF formalism provides a flexible and scalable approach for developing diverse and specific force fields.