QM Dataset for MBIS Partial Charges: Training set for DASH partial charges

Abstract

This data set contains 341'811 molecules with elements from the organic set (C, H, N, O, P, S, Cl, Br, I, F) and a molecular weight of up to 500 g/mol with a large diversity in Morgan Fingerprints r=2 (MF2). For each molecule three conformations are available. MBIS charges were calculated using Psi4 with a TPSSh functional and a def2-TZVP basis set in a implicit solvent (chloroform) with the PCM model. The charges are saved for each conformer as a properties and are seperated by a "|" in the .sdf file.