amino3to1 in protein-protein complexes about biopandas

[rasbt]

Thanks for pointing this out @karlafej !
I haven't worked with multi-chain proteins in recent projects and completely forgot to include them in the test cases, which should definitely be addressed like you said. I just see that there's another problem in the current implementation since it assumes unique residue numbers, which is a bad assumption for multi-domain cases ... I will fix that :).

About the returned values from the amino3to1 function. I think a dictionary could be a good idea, like you suggested, but I would favor returning a list of string sequences to preserve the order in which the chains appear in the PDB flle.

For example, for 5mtn is would return

['SLEPEPWFFK...', 'SVSSVPTKLE...']

and the chain ideas could be obtained via

pdb.df['ATOM']['chain_id'].unique()

if desired. For instance, one could iterate

for sequence, chain_id in zip(amino3to1_results, pdb.df['ATOM']['chain_id'].unique()):
    # do something

Alternatively, amino3to1 could return a list of tuples

[('A', 'SLEPEPWFFK...'), ('B', 'SVSSVPTKLE...')]

Any thoughts?

[rasbt]

I just updated the amino3to1 method (in #22 ) to return a DataFrame containing a column for the chain id. I think that this would be the most universally useful way. Let me know if you have any feedback or suggestions on this.
(I added some usage examples here)

[karlafej]

Thank you!