Transferable high dimensional neural network potentials (HDNNP) have shown great promise as an avenue to increase the accuracy and domain of applicability of existing atomistic force fields for organic systems relevant to life science. We have previously reported such a potential (Schrödinger-ANI) that has broad coverage of druglike molecules. We extend that work here to cover ionic and zwitterionic druglike molecules expected to be relevant to drug discovery research activities. We report a novel HDNNP architechture, which we call QRNN, that predicts atomic charges and uses these charges as descriptors in an energy model which delivers conformational energies within chemical accuracy when measured against the reference theory it is trained to. Further, we find that delta learning based on a semi-empirical level of theory approximately halves the errors. We test the models on torsion energy profiles, relative conformational energies, geometric parameters and relative tautomer errors.
All geometries used to analyze errors statistics in the main text with energy labels for QeqNN, QeqNN-TB, QRNN, QRNN-TB, GFN2-xTB, PM7 and wB97X-D/6-31G* are available for download in the file supplementary_data_files.tar.gz. Each test set is stored in a separate directory and each set of conformers, rotamers or tautomers is stored in a separate json file. The format for the files is given in a README.format file in each directory.
Additional tables and figures showing model results for all models discussed in this work as well as details on training to the QM9 dataset.