Open questions about implementing non canonical amino acids

[matteagle2]

Hi,

first of all, thank you for this amazing software!

I am having trouble understanding the instructions for including non canonical amino acids. I will now try to explain my lack of understanding and hope you excuse my denseness:

So even though it is not stated in the description, I assume that the input pdb file must include the non canonical amino acid. And I also assume that this residue must have the same 3 letter code description as in the json file. My first question about this is, whether the residue must have a different chain name just like regular ligands or should it be the same as the protein (which would make sense since it is part of the protein). Next up, it is unclear to me whether the --predict_ligand or --ligand_file flags should be used in this instance (i prefer the --ligand_file flag due to entry 10 of the FAQ). If one of the flags is to be used should the chain name be the same as the proteins chain name?
And my last question (at least for now;)) is about the 'leaving' key in the json file. In my understanding this is meant for instances where the actual backbone of the residue is modified. So if only the sidechain is to be modified, every item in the list would be set to False. Is that correct or am I completely wrong about this?

I got the following error:

Traceback (most recent call last):
File "/agh/scratch/Boje/PLACER/run_PLACER.py", line 282, in
main(args)
File "/agh/scratch/Boje/PLACER/run_PLACER.py", line 196, in main
outputs = placer.run(placer_input_iter, args.nsamples)
File "/agh/scratch/Boje/PLACER/PLACER.py", line 177, in run
assert k not in self.mols().keys(), f"Residue {k} already in database, please choose a different name3."
AssertionError: Residue LIG already in database, please choose a different name3.

with this command line: /agh/scratch/Boje/placer_env/bin/python /agh/scratch/Boje/PLACER/run_PLACER.py -f /agh/scratch/Boje/PLACER/test_cases/julian/test.pdb --odir /agh/scratch/Boje/PLACER/test_cases/retro/ --rerank prmsd -n 100 --ignore_ligand_hydrogens --ligand_file LIG:/agh/scratch/Boje/PLACER/test_cases/retro/LIG.sdf --residue_json /agh/scratch/Boje/PLACER/test_cases/retro/output.json

i renamed the ligand to something else with the same result...

output.json

thanks in advance

[SongXia-NYU]

Hi I just encountered the same problem. After printing out self.mols().keys(), I found that there are over 30,000 predefined residues. Your defined residue names cannot be predefined. After several tries, "YYF" seems to work for me.