Improved Treatment of 1-4 interactions in Force Fields for Molecular Dynamics Simulations

Abstract

Traditional force fields commonly use a combination of bonded torsional terms and empirically scaled non-bonded interactions to capture 1-4 energies and forces of atoms separated by three bonds in a molecule. While this approach can yield accurate torsional energy barriers, it often leads to inaccurate forces and erroneous geometries, and creates an interdependence between dihedral terms and non-bonded interactions, complicating parameterization and reducing transferability. In this paper, we demonstrate that 1-4 interactions can be accurately modeled using only bonded coupling terms, eliminating the need for arbitrarily scaled non-bonded interactions altogether. Furthermore by leveraging the automated parameterization capabilities of the Q-Force toolkit, we efficiently determine the necessary coupling terms without the need for manual adjustment. Our approach is first validated on a range of small molecule systems, encompassing both flexible and rigid structures, and shows a significant improvement in force field accuracy, obtaining sub-kcal/mol mean absolute error for every molecule tested. We further extend the bonded-only model for 1-4 interactions to Amber ff14sb, CHARMM36, and OPLS-AA force fields to reproduce ab initio gas and implicit solvent $φ,ψ$ surfaces of alanine dipeptide.

Figures (4)

• Figure 1: Flexible and rigid molecular datasets used in this study. Molecules in the (a) flexible dataset contain a single fully rotatable torsion, whereas (b) the rigid set has molecules with multiple torsions that are restricted due to double bonds, conjugation or being in a ring.
• Figure 2: Absolute energy errors (a) and the norm of the vectorial force error per atom (b) with both datasets combined for all benchmarks in the study. The bin widths are 0.25 kcal/mol for energies and 1.5 kcal/mol/$\AA$ for forces. The occurrences are normalized to be 1 at maximum occurrence for each method.
• Figure 3: Torsional energy profiles corresponding to the molecules in the flexible dataset for each method. Geometries are relaxed for each method to their constrained minimum.
• Figure 4: Ramachandran $\phi/\psi$ potential energy surfaces for alanine dipeptide. (A) $\omega$-B97X-V/def2-TZVPD, Q-Force for bonded and with (B) ff14sb 1-5+ nonbonded interactions, (C) CHARMM36 nonbonded 1-5+ interactions, (D) OPLS-AA nonbonded 1-5+ interactions. (E) $\omega$-B97X-V/def2-TZVPD with SMD and (F) Q-Force with ff14sb nonbonded interactions and GBN2 implicit solvent; energies are relative to each surface’s minimum to enable the comparison.