Effective Fragment Potentials for Flexible Molecules: Transferability of Parameters and Amino Acid Database

Area of Science:

• Computational chemistry
• Molecular modeling
• Biophysics

Background:

• Accurate description of noncovalent interactions is crucial for predictive modeling.
• The effective fragment potential (EFP) method decomposes interactions but requires recomputation for geometry changes.
• Recomputing EFP parameters creates computational and human bottlenecks, reducing simulation efficiency.

Purpose of the Study:

• To introduce and explore a "flexible EFP" approach for adjusting fragment parameters to different geometries.
• To reduce the computational cost and human effort associated with standard EFP simulations.
• To develop an automated protocol for flexible EFP simulations in large molecular systems like proteins.

Main Methods:

• Developed a flexible EFP protocol involving parameter adjustment based on translations and rotations of local atomic coordinate frames.
• Validated the protocol using amino acid dimers from cryptochrome protein molecular dynamics snapshots.
• Created a parameter database for standard amino acids to enable automated flexible EFP simulations.

Main Results:

• Flexible EFP achieved results in close agreement with the standard EFP procedure.
• Demonstrated significant reduction in computational cost compared to standard EFP.
• Successfully applied flexible EFP to compute binding and electronic energies for a lumiflavin chromophore in a protein.

Conclusions:

• Flexible EFP offers a computationally efficient and accurate alternative to standard EFP for molecular simulations.
• The developed parameter database and protocol facilitate large-scale protein simulations.
• This approach enhances the applicability of EFP for predictive modeling in biological and materials science.