Curated SPICE 2 Dataset:

Full dataset, version "full_dataset_v1":

This provides a curated hdf5 file for the SPICE 2 dataset (release v2.0.1) designed to be compatible with modelforge, an infrastructure to implement and train NNPs.  This datafile  includes 2,008,126 total conformers for 113,985 unique molecules.

When applicable, the units of properties are provided in the datafile,  encoded as strings compatible with the openff-units package.  For more information about the structure of the data file, please see the following:

Changes: In this version, for each record `total_charge` is stored as an array of shape (N_conformers, 1), i.e., a value for each conformer; previously this was just a single value for each record as the charge state doesn't change for each conformer.  

When applicable, the units of properties are provided in the datafile,  encoded as strings compatible with the openff-units package.  For more information about the structure of the data file, please see the following:

• https://github.com/choderalab/modelforge/wiki/Dataset-and-curation#curation-module

This curated dataset was generated using the modelforge software at commit <add commit>:

• Link to the source code at this commit: <add commit>
• Link to the script file used to generate the dataset: <add commit>

Source Dataset:

Small-molecule/Protein Interaction Chemical Energies (SPICE).

The SPICE 2 dataset contains roughly 2 million conformations for a diverse set of small molecules, dimers, dipeptides, and solvated amino acids. It includes 17 elements, charged and  uncharged molecules, and a wide range of covalent and non-covalent interactions.  SPICE 2 is an update to spice 1, roughly double the total dataset size and including 2 additional elements.  It provides both forces and energies calculated at the ωB97M-D3(BJ)/def2-TZVPPD level of theory, using Psi4 1.4.1 along with other useful quantities such as multipole moments and bond orders.

Citations:

Original SPICE 1 publication:

• Eastman, P., Behara, P.K., Dotson, D.L. et al. SPICE,  A Dataset of Drug-like Molecules and Peptides for Training Machine Learning Potentials. Sci Data 10, 11 (2023). https://doi.org/10.1038/s41597-022-01882-6

Source dataset, released with CCO 1.0 Universal license:

• Eastman, P., Behara, P. K., Dotson, D., Galvelis, R., Herr, J., Horton, J., Mao, Y., Chodera, J., Pritchard, B., Wang, Y., De Fabritiis, G., & Markland, T. (2024). SPICE 2.0.1 (2.0.1) [Data set]. Zenodo. https://doi.org/10.5281/zenodo.10975225