Using PC clusters to evaluate the transferability of molecular mechanics force fields for proteins

Area of Science:

• Computational Chemistry
• Biophysics
• Molecular Modeling

Background:

• Assessing molecular mechanics parameter transferability from small molecules to proteins is challenging.
• Molecular dynamics simulations often lack sufficient length to explore diverse conformations for peptides.

Purpose of the Study:

• To develop a rapid method for evaluating molecular mechanics force field properties using large conformation sets.
• To verify suspected biases in AMBER force fields (ff94, ff99) and propose improvements.

Main Methods:

• Utilized a PC cluster to generate extensive native and non-native peptide conformations.
• Applied these atomic-detail conformation sets to evaluate force field performance and identify biases.

Main Results:

• Confirmed a bias towards alpha-helical structures in ff94 and ff99 AMBER force fields.
• Modified the ff99 force field based on evaluation results, leading to improved peptide stability and folding simulations.

Conclusions:

• The generated conformation sets enable rapid force field evaluation and modification.
• The improved force field demonstrates more accurate structural behavior in molecular dynamics simulations.
• The methodology offers transferable insights for future force field development and validation.