Harmonic Force Constants for Molecular Mechanics Force Fields via Hessian Matrix Projection

Area of Science:

• Computational Chemistry
• Molecular Modeling
• Quantum Chemistry

Background:

• Molecular mechanics force fields require accurate parameters for bond and angle properties.
• Existing methods for parameter derivation can be complex and interdependent.
• The Seminario method offers a route from quantum mechanics to force field parameters.

Purpose of the Study:

• To propose a modification to the Seminario method for deriving accurate harmonic bond and angle molecular mechanics force field parameters.
• To automate the parameter derivation process directly from quantum mechanical data.
• To provide a comprehensive set of parameters for amino acids to aid protein force field development.

Main Methods:

• Modification of the Seminario method to utilize the quantum mechanical Hessian matrix.
• Direct computation of bond and angle parameters from quantum mechanical data.
• Automation of the entire parameter calculation process.

Main Results:

• Reduced average error in reproducing quantum mechanical normal-mode frequencies from 12.3% to 6.3% for 70 benchmark molecules.
• Achieved full automation of the parameter derivation process.
• Eliminated interdependency between bond, angle, and other force field parameters.

Conclusions:

• The modified Seminario method provides a more accurate and automated approach to deriving molecular mechanics force field parameters.
• This method enhances the reliability of parameters used in molecular simulations.
• The provided amino acid parameters will facilitate future advancements in protein force field development.