ValueError: No chains with given name found

[b-schubert]

I encountered the following error:

Traceback (most recent call last):
  File "/Users/bs224/Dropbox/PostDoc/projects/EVcouplings/test/TestPDB.py", line 33, in test_pbd
    distmap_intra = intra_dists(selected_structures)
  File "/Users/bs224/Dropbox/PostDoc/projects/EVcouplings/evcouplings/compare/distances.py", line 759, in intra_dists
    model
  File "/Users/bs224/Dropbox/PostDoc/projects/EVcouplings/evcouplings/compare/distances.py", line 667, in _prepare_chain
    chain = structures[pdb_id].get_chain(pdb_chain, model)
  File "/Users/bs224/Dropbox/PostDoc/projects/EVcouplings/evcouplings/compare/pdb.py", line 538, in get_chain
    "No chains with given name found"
ValueError: No chains with given name found

which basically breaks the complete pipeline. The culprit was 1OAY...

I propose to ignore the particular PDB structure and continue with the processing via a try and except block. Agreed?

[b-schubert]

See PR #50

[thomashopf]

Can you please describe in a bit more detail in what setting the problem happened? Basically it is trying to access a chain in the structure that the SIFTS mapping claims is there, but is not actually present in the MMTF structure?

• What chain is it?
• Are the SIFTS mapping files up to date or is it still the ones I made in April? Or is there actually a mismatch between SIFTS and what is in the PDB?

I wouldn't be too happy just to ignore the error unless absolutely necessary.

[b-schubert]

• its chain I
• happened during alignment-based comparison.
• PDB has the chain annotated for 1OAY (http://www.ebi.ac.uk/pdbe/entry/pdb/1oar/protein/1)
• SIFTS file is the old
• MMTF does not contain the chain

  print(self.mmtf.chain_name_list)
  >['H', 'J', 'L', 'M', 'N', 'O', 'H', 'J', 'H', 'J', 'L', 'M', 'N', 'O']

Updating the SIFTS file and fasta solved the discrepancy. I suggest I'll update the SIFTS file on orchestra.

[thomashopf]

Okay great that this solves it, the files were a couple of months old already I think....