Error estimates for (semi-)empirical dispersion terms and large biomacromolecules

Area of Science:

• Computational chemistry and materials science
• Biomolecular modeling and simulation
• Quantum mechanics and condensed matter physics

Background:

• First-principles modeling of biomaterials has advanced significantly due to increased computing power and density functional theory (DFT) applications.
• Dispersion corrections, particularly van der Waals (vdW) interactions, are vital for accurate modeling but have limited error estimates for large biomolecules.

Purpose of the Study:

• To derive error estimates for dispersion terms in DFT, semiempirical QM (SQM), and molecular mechanics (MM) methods.
• To assess the impact of these dispersion terms on the accuracy of biomolecular simulations, especially for large systems.

Main Methods:

• Utilized error statistics from smaller systems to estimate errors in dispersion terms.
• Analyzed dispersion contributions from the PDBbind database of protein-ligand interactions.
• Evaluated dispersion approximations used in DFT, SQM, and MM methods.

Main Results:

• Dispersion terms in DFT, SQM, and MM methods were analyzed for accuracy in large biomolecular systems.
• Error estimates indicate that dispersion terms are typically not the primary limitation for achieving chemical accuracy.
• Certain force fields and large ligand sizes present challenges for accurate dispersion modeling.

Conclusions:

• Dispersion corrections are generally reliable for biomolecular modeling, contributing to overall accuracy.
• While usually not a limiting factor, specific force fields and ligand complexities require careful consideration.
• The findings provide valuable insights for selecting appropriate computational methods for large biomolecular systems.