The details in generating <PDBid>_nowat.pdb

[SanFran-Me]

In the PDBbind2016 data directory, I see the following files that seems to be used in your training project.
<PDBid>_nowat.pdb -- Receptor structure with all HETATOMS removed

I was wondering how to generate thenowat.pdbforpocket.pdbfiles in PDBbind 2019 dataset, and the ChatGPT told me to use:
for file in *.pdb; do gnina -i "$file" -o "${file%.pdb}_nowat.pdb" --autobox_ligand ""; done

I am not sure if this is the same way of your project. If so, even though it works, it might have some differences with your dataset. So I am here to ask your methods to generate 'nowat.pdb' in details.

[dkoes]

You generate a pdb file without HETATM records by removing all the HETATM records from the file. This is not something gnina does. You can use your favorite cheminformatic toolkit or just run grep -v HETATM.

[SanFran-Me]

You generate a pdb file without HETATM records by removing all the HETATM records from the file. This is not something gnina does. You can use your favorite cheminformatic toolkit or just run grep -v HETATM.

Yeah, grep -v HETATM works. But some amino acids modified by phosphorylation will also be deleted（like T[THR:phosphorylated]）, how to avoid these occasions when deleting waters and ions in .pdb file ?

[dkoes]

Use your favorite cheminformatic tool. PyMOL, ProDY, etc can all do this sort of thing. This is outside the scope of gnina.