This is a test version of Biostars. For the public version, visit https://www.biostars.org.
Biopython: work with loops and qualifiers from multiple records

I have a gbk of a draft bacterial genomes (meaning multiple contigs). I want to extract some info about specific genes, and I get an error I cannot understand.

Parse the gbk

record_dict = SeqIO.to_dict(SeqIO.parse("SpeciesA.gbk", "genbank"))

My file looks:

    for i in record_dict:
    ...     for f in record_dict[i].features:
    ...             print f.qualifiers
{'locus_tag': ['AA_03640'], 'gene': ['ftsH_4']}
{'locus_tag': ['AA_03640'], 'inference': ['ab initio prediction:Prodigal:2.6', 'similar to AA sequence:UniProtKB:P37476'], 'codon_start': ['1'], 'EC_number': ['3.4.24.-'], 'transl_table': ['11'], 'product': ['ATP-dependent zinc metalloprotease FtsH'], ........

Now if I try to get out the contig name where the gene AA_03640 is in, as follow:

for i in record_dict:
...     for f in record_dict[i].features:
...             if f.qualifiers['locus_tag'] == 'AA_03640':
                    print(i)

But I get the following error:

Traceback (most recent call last):
  File "<stdin>", line 3, in <module>
KeyError: 'locus_tag'

Any help to understand where I am wrong? Thanks

genome python gene software error biopython
• 4,717 views • link
updated by Peter Peter • written by dago dago

The error is quite self-explanatory I would say, and means that f.qualifiers doesn't have a key "locus_tag"...

You can check which keys it does have using f.qualifiers.keys()

• 0 views • link
written by WouterDeCoster WouterDeCoster

I thought the same, but when I check the keys locus_tag is in there:

['locus_tag', 'gene']
['locus_tag', 'inference', 'codon_start', 'EC_number', 'transl_table', 'product', 'translation', 'gene', 'protein_id']
• 0 views • link
written by dago dago

Odd. For a moment I thought this was an easy question ;-)

Maybe use a try-except block:

try:
    if f.qualifiers['locus_tag'] == 'AA_03640':
        #stuff
except KeyError:
    print(f.qualifiers)

An unrelated problem is that you should check for f.qualifiers['locus_tag'] == ['AA_03640'], since it's a list and not a character.

• 0 views • link
written by WouterDeCoster WouterDeCoster

ok, I am officially confused. If I run:

for i in record_dict:
    for f in record_dict[i].features:
        print f.qualifiers.keys()

I get 

['locus_tag', 'inference', 'codon_start', 'product', 'transl_table', 'translation', 'protein_id']
['locus_tag']
['locus_tag', 'inference', 'codon_start', 'product', 'transl_table', 'translation', 'protein_id']
['locus_tag']

But If I run

for i in record_dict:
    for f in record_dict[i].features:
        try:
            if f.qualifiers['locus_tag'] == ['AA_03640']:
                print i
        except KeyError:
            print(f.qualifiers.keys())

I get

['estimated_length', 'linkage_evidence', 'gap_type']
['strain', 'mol_type', 'organism']
['strain', 'mol_type', 'organism']

I should have the same keys, shouldn't I?

• 0 views • link
written by dago dago

I am officially confused

I join being confused.

I don't really have experience with genbank files so I can't really nail this down. Maybe something for Peter, although I'm not sure if he visits biostars often...

• 0 views • link
written by WouterDeCoster WouterDeCoster

Double check your input genbank isn't malformed and that its /locus_tag fields are correct?

If you hit Peter up on twitter he's been really helpful to me in the past (and hes always on twitter ;) )

• 0 views • link
written by Joe Joe

I checked the gbk, it looks fine. I will try twitter..:)

• 0 views • link
written by dago dago

They were faster then me!=)

• 0 views • link
written by dago dago

To answer your question, no, you should not assume all the features have the same keys. It varies dramatically by the feature type - the source feature (usually there is only one) has things like the organism, while while CDS and gene features have things like a locus tag, and the CDS often has a translation. You probably want to add if f.type == "CDS": or if f.type == "gene": to your loop. See also http://www.warwick.ac.uk/go/peter_cock/python/genbank which does something similar to what I think you want to achieve.

(I have expanded on this as a full answer)

• 0 views • link
written by Peter Peter

1 answer

Not all feature types will have a locus tag - you probably only care about CDS or gene features. Even then, not genes will have a locus tag (although generally within a record they either all do, or all don't).

Finally if there is a locus tag, then f.qualifiers['locus_tag'] will return a list (since in general the annotation qualifiers can appear more than once), so you'd want to check using 'AA_03640' in f.qualifiers['locus_tag'] or similar, e.g.

for i in record_dict:
    for f in record_dict[i].features:
        if f.type == "CDS":
            if 'AA_03640' in f.qualifiers.get('locus_tag', []):
                print(i)

See also http://www.warwick.ac.uk/go/peter_cock/python/genbank which does something similar to what I think you want to achieve.

• 0 views • link
written by Peter Peter

Hi, thank you very much for the answer! You are right! Adding if f.type == "CDS" fixed it!

• 0 views • link
written by dago dago

Log in to answer this question.

More posts like this