Better force fields start with better data: A data set of cation dipeptide interactions

Abstract

We present a data set from a first-principles study of amino-methylated and acetylated (capped) dipeptides of the 20 proteinogenic amino acids – including alternative possible side chain protonation states and their interactions with selected divalent cations (Ca2+, Mg2+ and Ba2+). The data covers 21,909 stationary points on the respective potential-energy surfaces in a wide relative energy range of up to 4 eV (390 kJ/mol). Relevant properties of interest, like partial charges, were derived for the conformers. The motivation was to provide a solid data basis for force field parameterization and further applications like machine learning or benchmarking. In particular the process of creating all this data on the same first-principles footing, i.e. density-functional theory calculations employing the generalized gradient approximation with a van der Waals correction, makes this data suitable for first principles data-driven force field development. To make the data accessible across domain borders and to machines, we formalized the metadata in an ontology.

System-Specific Parameter Optimization for Nonpolarizable and Polarizable Force Fields

“…The reference data can be found on the NOMAD repository via the DOI: 10.17172/NOMAD/2023.02.03-1. 72 FFAFFURR can be found at: https://github.com/XiaojuanHu/ffaffurr-dev/ releases/tag/version1.0. 97 Code for implementing CTPOL in OpenMM: https://github.com/XiaojuanHu/CTPOL_MD.…”

#COVIDisAirborne: AI-enabled multiscale computational microscopy of delta SARS-CoV-2 in a respiratory aerosol

“…In contrast, OrbNet allows for DFT accuracy with over 1000-fold speedup, providing a useful method for benchmarking and refining the force field simulation parameters with quantum accuracy ( Christensen et al, 2021 ). To confirm the accuracy of OrbNet versus DFT ( ω B97X-D/def2-TZVP), the inset of Figure 4(E) correlates the two methods for the Ca 2+ -binding energy in a benchmark dataset of small Ca 2+ -peptide complexes ( Hu et al, 2021 ). The excellent correlation of OrbNet and DFT for the present use case is clear from the inset figure; six datapoints were removed from this plot on the basis of a diagnostic applied to the semi-empirical GFN-xTB solution used for feature generation of OrbNet ( Christensen et al, 2021 ).…”

“…E) OrbNet Sky energies versus CHARMM36m energies for each sub-selected system, colored by total number of atoms. Performance of OrbNet Sky versus DFT in subplot ( ω B97x-D3/def-TZVP, R 2 =0.99, for 17 systems of peptides chelating Ca 2+ (Hu et al, 2021)). Visualized with VMD.…”

Potential Energy Surfaces Sampled in Cremer–Pople Coordinates and Represented by Common Force Field Functionals for Small Cyclic Molecules

“…Sampling the conformational space via dihedral angles is a common practice. − However, this method is not efficient for exploring the conformational space of small cyclic molecules and produces ambiguous samplings due to the redundancy of internal coordinates. As a result, only a portion of the PES is sampled, and the poor representations of the conformational space often require further refinements of the torsion parameters. ,,, More importantly, the dihedral-based sampling fails to account for the underlying physics ,,− of conformational transitions, a feature critical − for predicting solvation properties. This problem is more pronounced in the development of machine-learned force fields (FFs), where adequately sampling the PESs is crucial. − …”