the protein pdb file need to be fixed

[wangjx22]

Dear authors, I have an issue when running the script "python3.6 ../PLANET_run.py -p adrb2.pdb -l adrb2_ligand.sdf -m mols.sdf". The error is "the protein pdb file need to be fixed" in line 21 of PLANET_run.py Could you help me to solve this problem.

yours, Jinxian

[aquilazhang]

Hello, Jinxian: This issue happens when the protein structure (.pdb) is incomplete (missing alpha-carbon atoms). You need to prepare the protein structure with other third-party software.

[wangjx22]

Thanks for your reply. Could you send me complete data like a protein or a molecule to run the code successfully? My email is wangjx22@m.fudan.edu.cn.

yours, Jinxian

[aquilazhang]

The files in 'demo' folder can be run successfully. Please provide a full traceback to let me see what happens.

[wangjx22]

I use the demo to run the script "python3.6 ../PLANET_run.py -p adrb2.pdb -l adrb2_ligand.sdf -m mols.sdf". But have an error: Traceback (most recent call last): File "/remote-home/jinxianwang/code/PLANET-main/PLANET_run.py", line 19, in set_pocket_from_ligand self.pocket = ProteinPocket(protein_pdb=protein_pdb,ligand_sdf=ligand_sdf) File "/remote-home/jinxianwang/code/PLANET-main/chemutils.py", line 95, in init ligand = Mol(Chem.SDMolSupplier(ligand_sdf,sanitize=False)[0]) OSError: File error: Bad input file adrb2_ligand.sdf

During handling of the above exception, another exception occurred:

Traceback (most recent call last): File "/remote-home/jinxianwang/code/PLANET-main/PLANET_run.py", line 174, in predicted_affinities,mol_names,smis = workflow(protein_pdb,mol_file,ligand_sdf,centeriod_x,centeriod_y,centeriod_z) File "/remote-home/jinxianwang/code/PLANET-main/PLANET_run.py", line 107, in workflow estimator.set_pocket_from_ligand(protein_pdb,ligand_sdf) File "/remote-home/jinxianwang/code/PLANET-main/PLANET_run.py", line 21, in set_pocket_from_ligand raise RuntimeError('the protein pdb file need to be fixed') RuntimeError: the protein pdb file need to be fixed

[aquilazhang]

Did you enter the 'demo' dir before running the above command? Or you can run the script with absolute path of required files.

[wangjx22]

Thanks for your help. I can run the code now, however, I want to test new molecules. Which .py file could transform molecules into an SDF file?

Yours Jinxian

[aquilazhang]

There is no script for transforming molecules into an SDF file in this repository. The transformation is usually performed with OpenBabel or other softwares for molecule modeling.

[wangjx22]

I have noticed that class VS_SMI_Dataset in PLANET_run.py. So, we also can input the SMILES of molecules? Is it possible to provide input file format in SMILES?

[aquilazhang]

Yes, that's right~ Molecules in SMILES format also works

[wangjx22]

I have tested my protein, but have 2 errors. the first is Traceback (most recent call last): File "/remote-home/jinxianwang/code/PLANET-main/demo/../PLANET_run.py", line 19, in set_pocket_from_ligand self.pocket = ProteinPocket(protein_pdb=protein_pdb,ligand_sdf=ligand_sdf) File "/remote-home/jinxianwang/code/PLANET-main/chemutils.py", line 100, in init self.res_features = torch.from_numpy(np.concatenate([residue.get_feature() for residue in self.pocket_residues],axis=0)) File "<__array_function__ internals>", line 5, in concatenate ValueError: need at least one array to concatenate

The second is Traceback (most recent call last): File "/remote-home/jinxianwang/code/PLANET-main/demo/../PLANET_run.py", line 182, in predicted_affinities,mol_names,smis = workflow(protein_pdb,mol_file,ligand_sdf,centeriod_x,centeriod_y,centeriod_z) File "/remote-home/jinxianwang/code/PLANET-main/demo/../PLANET_run.py", line 107, in workflow estimator.set_pocket_from_ligand(protein_pdb,ligand_sdf) File "/remote-home/jinxianwang/code/PLANET-main/demo/../PLANET_run.py", line 21, in set_pocket_from_ligand raise RuntimeError('the protein pdb file need to be fixed') RuntimeError: the protein pdb file need to be fixed

I have pasted the protein file 4kcd.pdb.gz , could you help me to check what problem in my file.

[aquilazhang]

I have prepared the protein structure, you can try it out. (unzip it first 4kcd_prepared.zip )

[wangjx22]

thanks for your help. But, I have the same errors using the prepared protein.

[aquilazhang]

Did you unzip the file? Besides, since you may have trouble in transforming a molecule file into SDF, how did you assign the flag "-l"? Or, you may use the "-x" "-y" "-z" instead of "-l". But I am not sure if you have defined the pocket center correctly.

[wangjx22]

I am sure unzip the file. I using the 4kcd_prepared.pdb you given to me and mols.sdf in your demo as the input of model. And I did not assign the fiag "-I". I use the script: python ../PLANET_run.py -p 4kcd_prepared.pdb -m mols.sdf

[aquilazhang]

The pocket center is required by PLANET to extract pocket residues from given PDB structure, which is defined as the mass center of molecule through "-l" flag. Or, specify the coordinate of pocket center through "-x" "-y" "-z".

[wangjx22]

The adrb2_ligand.sdf is necessary? I found adrb2_ligand.sdf is a crystal ligand file for determining binding pocket. If a protein have multi binding pocket, so the adrb2_ligand.sdf should include multi molecules?

[aquilazhang]

It is necessary. The SDF file to define pocket should contain only ONE molecule. For multi-pockets, you need to specify one pocket and run PLANET at a time and do this for several rounds .