Docking computations failed for certain complexes

[RafiBrent]

Thanks so much for creating this tool! I've largely had success in using the published Colab notebook-- however, for a certain subset of complexes DiffDock eventually prints "Failed for 1 complexes" and then moves on without throwing an exception or producing any output. I have confirmed that the HuggingFace app also encounters an error, so this issue is likely not an artifact of the Colab notebook itself. A reproducible example with this behavior is PDB ID 6seo with the following ligand SMILES string (obtained through RDKit interpretation of the associated mol2 file from PDBbind):

CC(=O)N[C@H](C(=O)N1CCC[C@H]1C(=O)N[C@@H](CC(C)C)C(=O)N[C@@H](CCCNC(N)=[NH2+])C(=O)N[C@@H](C)C(=O)N[C@@H](CCCNC(N)=[NH2+])C(=O)N[C@@H](CC(N)=O)C(=O)N[C@@H](CC(C)C)C(=O)N1CCC[C@H]1C(=O)N1CCC[C@H]1C(=O)N[C@@H](CO)C(=O)N[C@@H](Cc1ccccc1)C(=O)N[C@@H](Cc1ccccc1)C(=O)N[C@H](C(=O)N[C@@H](CCC(=O)[O-])C(=O)N1CCC[C@H]1C(N)=O)[C@@H](C)O)C(C)C

Please let me know if you have any advice regarding this issue and I sincerely appreciate your help.

[gcorso]

Hi @RafiBrent, if running the DiffDock code directly you should also be able to see the specific error that occurs with that specific complex. Given the relatively complex ligand, I expect that the issue is that RDKit has issues providing initial structures. Maybe you could try using the command --keep_local_structures to avoid these RDKit embeddings. I hope this helps!