Fast Fourier Transform Enables Automated Parametrization of Complex Dihedral Potentials in All-Atom and...

Area of Science:

• Computational Chemistry
• Molecular Modeling
• Force Field Development

Background:

• Torsional parametrization is a significant challenge in modern molecular force fields.
• Existing methods struggle with asymmetric and multimodal dihedral energy profiles.

Purpose of the Study:

• To develop a general and automated methodology for dihedral parametrization.
• To address limitations of traditional methods for complex torsional profiles.

Main Methods:

• Utilized Fast Fourier Transform (FFT) analysis for dihedral parametrization.
• Combined FFT with QM-MM energy matching (AA) and Iterative Boltzmann Inversion (CG) for refinement.

Main Results:

• Developed a robust FFT-based framework for reconstructing complex torsional energy profiles.
• Achieved quantitative reproduction of reference energy landscapes in diverse chemical environments.
• Demonstrated accurate modeling of conformational behavior for MS-Z and Aβ42 peptide.

Conclusions:

• FFT-based torsional parametrization offers a systematic and unbiased approach.
• This method enables the development of next-generation molecular force fields with improved accuracy.
• The methodology is applicable to both all-atom and coarse-grained models.