Error in saliency maps

[ga01]

Dear DeepFRI developers,
There seems to be an error in the saliency maps when the latest release and the downloadable models are used. For example in the case of 2PE5 in the Nat. Comms paper for the DNA-binding GO term (GO:0003677) the DNA-binding domain of the protein is activated (~1-50 residues, Figure 4C), but with the current version of the tool and models, residues that are not involved in DNA-binding are activated:

Best wishes,
George

[ga01]

PS.: It seems that when using the new models, all three proteins (2w83-E, 2J9H-A, 2PE5-B) of Figure 4C of the DeepFRI paper (https://doi.org/10.1038/s41467-021-23303-9) have saliency maps that do not match the pattern on Figure 4C, and also does not correspond with the ligand binding residues.

I wonder whether it would be possible to provide also the original models that were used in the paper for download?

[MiJia-ID]

Excuse me, but how are the activation sites of BioLip added to the graph? Where is the data where to download it?

[MiJia-ID]

[Young0222]

Excuse me, but how are the activation sites of BioLip added to the graph? Where is the data where to download it?

I also want to ask this question, can anyone provide an answer?