Data Sources and Attribution
This release packages processed protein objects, labels, and split lists for E3former. Source data retain their original notices and license terms. Our packaging and documentation are licensed under CC BY 4.0.
| Component | Source | Upstream license |
|---|---|---|
| CATH classification | CATH | CC BY 4.0 |
| PPBS labels and evaluation subsets | ScanNet | Apache 2.0 |
The upstream ScanNet license is retained unchanged. The processed objects in this release were prepared for the E3former data loader; they are not the original ScanNet distribution.
ScanNet reference: Tubiana, J., Schneidman-Duhovny, D., and Wolfson, H. J. ScanNet: an interpretable geometric deep learning model for structure-based protein binding site prediction. Nature Methods 19, 730-739 (2022). https://doi.org/10.1038/s41592-022-01490-7
AlphaFold reference: Jumper, J. et al. Highly accurate protein structure prediction with AlphaFold. Nature 596, 583-589 (2021). https://doi.org/10.1038/s41586-021-03819-2
The E3former code license does not replace upstream dataset licenses.