search

HannesStark / EquiBind

EquiBind: geometric deep learning for fast predictions of the 3D structure in which a small molecule binds to a protein
MIT License
473 stars 109 forks source link

Problem with pdb file input #7

Closed AlexExplorer1 closed 2 years ago

AlexExplorer1 commented 2 years ago

Preprocessed pdb file as suggested in arxiv. Trying to make the prediction but got an error:

Traceback (most recent call last): File "inference.py", line 471, in inference_from_files(args) File "inference.py", line 339, in inference_from_files rec, rec_coords, c_alpha_coords, n_coords, c_coords = get_receptor(rec_path, lig, cutoff=dp['chain_radius']) File "E:\EquiBind-main\commons\process_mols.py", line 373, in get_receptor c_alpha_coords = np.concatenate(valid_c_alpha_coords, axis=0) # [n_residues, 3] File "<__array_function__ internals>", line 6, in concatenate ValueError: need at least one array to concatenate

HannesStark commented 2 years ago

Hi, Are you using the data as we provide it on zenodo?

To figure out what is going on, I would recommend printing the name of the complex for which this is happening and looking at how many chains it has and how big the chain is in which this issue is occuring.

AlexExplorer1 commented 2 years ago

Hi, Thank you for answer! I am using pdb files from https://www.rcsb.org/ I was trying the different proteins, but got pretty much the same error. Here is example: 2SBL, has 2 chains and length of 839 Is any restriction for the number of chains or length of protein?

HannesStark commented 2 years ago

There should be no restriction on the number of chains or the length of the protein. What are the lengths of the chains in your example. And maybe it helps to visualize the complex to figure out what is going on. I would use PyMol for that.

AlexExplorer1 commented 2 years ago

Thank you. I solved problem using this suggestion: https://github.com/HannesStark/EquiBind/issues/9