Hi there,

Great work on PocketGen! I've been reviewing your paper and code, and I have a question about the data splitting and validation process for the CrossDocked dataset.

I downloaded the crossdocked_split.pt file from Zenodo and noticed something unexpected:

a = torch.load('crossdocked_split.pt')
(set(a['val']) & set(a['test'])) == set(a['test'])

This returns True, indicating that the validation and test sets are identical. Is this intentional?

Could you please clarify if this is the intended data split, and if so, the reasoning behind it?

Thanks for any clarification you can provide!

Hello, I encountered some issues while trying to run your model with the use case '2p16'.
Here is the detailed error message:

Traceback (most recent call last):
File "generate_new.py", line 305, in
aar, rmsd, attend_logits = model.generate(batch, dir_name)
File "/home/dell/download/pocketgen/models/PD.py", line 354, in generate
res_H, res_X, pred_ligand, ligand_feat, pred_res_type, attend_logits = self.encoder(res_H, res_X, res_S, res_batch, pred_ligand, ligand_feat, batch['ligand_mask'], batch['edit_residue_num'], residue_mask)
ValueError: not enough values to unpack (expected 6, got 5)

It seems that the return values from the function call are not as expected. Can you provide guidance on how to resolve this issue?
Thank you for your assistance!

Very interesting work!

Although your method is trained to predict the full pocket, do you expect it to work
properly (if not better) for partial binding site impainting (Fixing a parte of the residues)?
Are you planning to implement such possibility in the scripts or to fine-tune a model for such partial pocket sequence retrieval?

Thanks!

Hi,

Thanks for your work on this interesting model!

I'm trying create an environment using the yaml file provided but it seems Pip can't find suitable versions for many of the libraries needed, an excerpt from my error message;

Pip subprocess error:

ERROR: Ignored the following versions that require a different python version: 0.0.1 Requires-Python >=3.9,<4.0; 0.21 Requires-Python >=3.9; 0.21.1 Requires-Python >=3.9; 0.21.2 Requires-Python >=3.9; 0.21.post1 Requires-Python >=3.9; 0.21rc1 Requires-Python >=3.9;

ERROR: Could not find a version that satisfies the requirement pyg-lib==0.2.0 (from versions: none)
ERROR: No matching distribution found for pyg-lib==0.2.0

Is there a different version of the environment file that I can try to use?

Thanks so much for your help!