Dataset and predictions aviability

[podleyan]

Hello, happy new year!

very nice work on the paper! You did such a massive work creating dataset and comparing all this methods.

The dataset you used for training the models looks really good and it would be very useful for my research.
I'd like to have LIGYSIS final dataset (before adaptation for specific models) per residue with information about which residue is binding and which is not.
I tried to create it myself with the files available here and at https://zenodo.org/records/13121414, but unfortunately some pickles are missing/ have other names and it's not very straightforward how to obtain it. Could you please support me with that?
Is it LIGSITE_pockets_DEF_TRANS.pkl?

Also i would appreciate to have predictions per residue for each model you compared in the paper. Is this is a correct file? ALL_PREDS_COMBINED_RIGHT_EXTENDED_DEF.pkl

It would be very useful if you would also include some information at the start of the each notebook about what exactly each notebook does.

Thank you a lot!
Best, Yana