chain id parameter for structure.io.pdbx.get_sequence #600
Conversation
'get_sequence' in the module 'biotite.structure.io.pdb' to return a dictionary mapping chain_id to the sequence
| sequences : list of Sequence or dict | ||
| If `chain_ids` is False, returns a list of protein and nucleotide | ||
| sequences for each entity. | ||
| If `chain_ids` is True, returns a dictionary where each key is a | ||
| chain ID and each value is the corresponding sequence. |
There was a problem hiding this comment.
I think we should note here that the chain IDs correspond to atom_site.auth_asym_id.
| for entity, strand_ids in enumerate(strand_ids): | ||
| for strand_id in strand_ids: | ||
| strand_ids_to_seq_dict[strand_id] = sequences[entity-1] |
There was a problem hiding this comment.
This lines enumerates, so the first tuple value is not the entity (ID) but the index (always starting at zero).
| for entity, strand_ids in enumerate(strand_ids): | |
| for strand_id in strand_ids: | |
| strand_ids_to_seq_dict[strand_id] = sequences[entity-1] | |
| for i, strand_ids in enumerate(strand_ids): | |
| for strand_id in strand_ids: | |
| strand_ids_to_seq_dict[strand_id] = sequences[i] |
|
Thanks for the PR. I would be in favor of dropping the Note that this would require a few small adjustments of that function call in the tests and documentation (searching for |
|
I agree with @padix-key. I think we should drop the |
with entity_poly.pdbx_strand_id as keys. Updated test_pdx.py test_get_sequence to reflect returning a dict instead of a list
|
I sent an additional commit with the suggested changes as I understand them. It is worth noting that there may be instances where entity_poly.pdbx_strand_id is not equivalent to atom_site.label_asym_id. The structure used in the test function (PDB:5UGO) is an example of this where the strand ID is "T" and the asym_id is "A". This may be an edge case and I am unsure if it will cause issues but its worth mentioning. |
|
Thanks for the changes. There are still remaining
The strand ID is not matching |
| sequence_dict = { | ||
| strand_id: sequence | ||
| for sequence, strand_ids in zip(sequences, strand_ids) | ||
| for strand_id in strand_ids | ||
| } |
There was a problem hiding this comment.
In case some converted sequence is None in the list comprehension above, sequences would have a different length than strand_ids.
|
Please ignore those commits. I am still learning git and did not think those commits would be sent to this pull request. |
|
No problem. You may also convert this PR to a Draft PR to further indicate that this is work in progress |
|
The last commit is the one I would like to merge if possible. |
|
Looks good to me, however, one small adjustment is necessary: With the merge of #552 |
Added a new parameter 'chain_id' to the function 'get_sequence' in the module 'biotite.structure.io.pdbx' to return a dictionary mapping chain_id to the sequence