Data for: "ANI neural network potentials for small molecule pKa prediction"

Data for small molecule pKa prediction models

The data herein includes trained models, DFT validation outputs and datasets used for the above work.

"Reference_DFT_Level_Data.zip" - This .zip contains the output ORCA files for each structure in the charged/uncharged and gaseous and aqueous phases. In the paper this is referred to as the DFT Global Minimum approach.

"Reference_Inputs_and_Structures.zip" - This .zip contains all files that resulted from a) initial geometry creation in Avogadro and b) the resulting CREST structures that resulted from DFT level and Avogadro level inputs.

"Datasets.zip" - Contains the datasets in .h5 file format, used to train each of the four model types.

"Trained_Models.zip" - Contains all trained models used within the paper.

"Additional_Data.zip" - Contains all extra claculation information for structures outwith test set.