tiwarylab / grasp Goto Github PK

Hi, it seems triple bonds are not accounted for, as I am getting this error. Is it possible?

"""
Traceback (most recent call last):
  File "/usr/local/lib/python3.7/site-packages/joblib/externals/loky/process_executor.py", line 463, in _process_worker
    r = call_item()
  File "/usr/local/lib/python3.7/site-packages/joblib/externals/loky/process_executor.py", line 291, in __call__
    return self.fn(*self.args, **self.kwargs)
  File "/usr/local/lib/python3.7/site-packages/joblib/parallel.py", line 590, in __call__
    for func, args, kwargs in self.items]
  File "/usr/local/lib/python3.7/site-packages/joblib/parallel.py", line 590, in <listcomp>
    for func, args, kwargs in self.items]
  File "/content/GrASP/parse_files.py", line 690, in process_production_set
    raise e
  File "/content/GrASP/parse_files.py", line 688, in process_production_set
    process_system(mol2_dir + structure_name, save_directory=f'{prepend}/{data_dir}', parse_ligands=False)
  File "/content/GrASP/featurize_protein.py", line 225, in process_system
    edge_attributes = {tuple(bond.atoms.ids):{"bond_type":bond_type_dict[bond.order]} for bond in protein.bonds}
  File "/content/GrASP/featurize_protein.py", line 225, in <dictcomp>
    edge_attributes = {tuple(bond.atoms.ids):{"bond_type":bond_type_dict[bond.order]} for bond in protein.bonds}
KeyError: '3'
"""

I think it goes back to this bit here:

# Leaving an extra bit to denote self loops
    bond_type_dict = {
        '1': [1,0,0,0,0,0],
        '2': [0,1,0,0,0,0],
        'ar':[0,0,1,0,0,0],
        'am':[0,0,0,1,0,0],
        'un':[0,0,0,0,1,0]  # Unkown Bond Type
    }

How can we retrieve the amino acids comprising the active site? Is there a method to obtain information on the amino acids in the binding site?

Thank you

Is GrASP compatible with macOS? I think it might not be as certain dependencies, most of them CUDA-related seem not to be compatible with this OS.

Does GrASP work with mmCIF format?

I am trying to install the two conda environments using the .yml files provided. But both the installation fails with different errors. The ob_env fails with some package incompatibility error and pytorch_env fails with a pip error. I have attached the screenshot for both. I am trying to install them in a Redhat linux system.

ob_env error:

pytorch_env error:

The number of cores is hard coded and if unchanged, the execution might get killed:

Perhaps it would be better if there was an argument to specify this, or warn the user on the README.

It happened to me, now edited the code to use 8 cores, and all good.

Thanks!

Many of the proteins on my dataset are crashing on the featurize_protein.py, with a KeyError. The hybridisation state of the atom is SP, but only SP2 and SP3 are accounted for in the dictionary. How could this be fixed? Thanks!

A structure crashed with this error: ERROR: THIS CODE IS KNOWN TO FAIL WHEN ATOMS ARE VIRTUALLYON TOP OF ONE ANOTHER. YOU SUPPLIED TWO ATOMS 0.000000APART. QUITTING NOW.

I think this is related to mdtraj: mdtraj/mdtraj#1300.

I think mdtraj just clashes here:

It still clashes.

I have pre-processed my structures, to split into chains, and as a result some metadata are lost, such us the unit cell dimensions'. I get this warning from MDAnalysis:

~/miniconda3/envs/ob/lib/python3.7/site-packages/MDAnalysis/coordinates/PDB.py:747: UserWarning: Unit cell dimensions not found. CRYST1 record set to unitary values.
  warnings.warn("Unit cell dimensions not found. "

I think it is just to extract the coordinates, so it should not have any effect on the GrASP featurisation of the atoms, should it?

Thanks!

Your tool seems very useful, nice work! Could you please add a license, e.g. MIT, to the git? Thank you so much!

Nice work! However, I noticed that there is currently no publicly available dataset for download. Would it be possible to make the dataset open and accessible?

I believe the digests are broken and don't work. They might need to be removed in order to facilitate the installation of the environments.

You might want to remove references to your local build machine and email.

Can't wait to try this out!