An improved dataset of force fields, electronic and physicochemical descriptors of metabolic substrates.

In silico prediction of xenobiotic metabolism is an important strategy to accelerate the drug discovery process, as candidate compounds often fail in clinical phases due to their poor pharmacokinetic profiles. Here we present Meta QM , a dataset of quantum-mechanical (QM) optimized metabolic substrates, including force field parameters, electronic and physicochemical properties. Meta QM comprises 2054 metabolic substrates extracted from the MetaQSAR database. We provide QM-optimized geometries, General Amber Force Field (FF) parameters for all studied molecules, and an extended set of structural and physicochemical descriptors as calculated by DFT and PM7 methods. The generated data can be used in different types of analysis. FF parameters can be applied to perform classical molecular mechanics calculations as exemplified by the validating molecular dynamics simulations reported here. The calculated descriptors can represent input features for developing improved predictive models for metabolism and drug design, as exemplified in this work. Finally, the QM-optimized molecular structures are valuable starting points for both ligand- and structure-based analyses such as pharmacophore mapping and docking simulations.