def get_hmmdb_ec_nums(acc, c):
"""
This returns a list of bioannotation:ECAnnotation objects
"""
qry = "SELECT he.ec_id FROM hmm_ec he JOIN hmm ON he.hmm_id=hmm.id WHERE hmm.accession = ?"
c.execute(qry, (acc, ))
ec_annots = list()
for row in c:
ec = annotation.ECAnnotation(number=row[0])
ec_annots.append(ec)
return ec_annots
def get_uniref_ec_nums(acc, c):
"""
This returns a list of bioannotation:ECAnnotation objects
"""
qry = """
SELECT us_ec.ec_num
FROM uniref_ec us_ec
WHERE us_ec.id = ?
"""
c.execute(qry, (acc, ))
ec_annots = list()
for row in c:
ec = annotation.ECAnnotation(number=row[0])
ec_annots.append(ec)
return ec_annots
def get_uspdb_ec_nums(acc, c):
"""
This returns a list of bioannotation:ECAnnotation objects
"""
qry = """
SELECT us_ec.ec_num
FROM uniprot_sprot_ec us_ec
JOIN uniprot_sprot_acc us_acc ON us_ec.id=us_acc.id
WHERE us_acc.accession = ?
"""
c.execute(qry, (acc, ))
ec_annots = list()
for row in c:
ec = annotation.ECAnnotation(number=row[0])
ec_annots.append(ec)
return ec_annots
def parse_kegg_blast_evidence(log_fh, polypeptides, blast_list, eval_cutoff):
'''
Reads a list file of NCBI BLAST evidence against KEGG and a dict of polypeptides,
populating each with Annotation evidence where appropriate. Only attaches evidence if
the product name is the default.
Currently only considers the top BLAST hit for each query which doesn't have
'uncharacterized' or hypothetical in the product name.
'''
for file in utils.read_list_file(blast_list):
last_qry_id = None
for line in open(file):
line = line.rstrip()
cols = line.split("\t")
# We're going to ignore any lines which have a few keywords in the name
# First character left off for initcap reasons
if 'ncharacterized' in cols[15] or 'ypothetical' in cols[15]:
continue
this_qry_id = cols[0]
# skip this line if it doesn't meet the cutoff
if float(cols[19]) > eval_cutoff:
continue
# the BLAST hits are sorted already with the top hit for each query first
if last_qry_id != this_qry_id:
annot = polypeptides[this_qry_id].annotation
# get the accession from the cols[5]
accession = cols[5]
# save it, unless the gene product name has already changed from the default
if annot.product_name == DEFAULT_PRODUCT_NAME:
accession = cols[5]
# the product field looks like this:
# dam; adenine-specific DNA methyltransferase; K06223 DNA adenine methylase [EC:2.1.1.72]
# troponin I type 1 (skeletal, slow); K10371 troponin I, slow skeletal muscle
if ' [EC' in cols[15] and cols[15].endswith(']'):
m = re.search("\; (K\d+)\s+(.+) \[EC\:(.+)\]",
cols[15])
else:
m = re.search("\; (K\d+)\s+(.+)", cols[15])
if m:
kegg_id = m.group(1)
product = m.group(2)
if len(m.groups()) == 3:
ec_num = m.group(3)
else:
ec_num = None
annot.product_name = product
log_fh.write(
"INFO: {0}: Updated product name to '{1}' based on BLAST hit to KEGG accession '{2}'\n"
.format(this_qry_id, annot.product_name,
accession))
if ec_num is not None and ec_num is not '':
ec = annotation.ECAnnotation(number=ec_num)
annot.add_ec_number(ec)
kegg_dbxref = annotation.Dbxref(db='KEGG',
identifier=kegg_id)
annot.add_dbxref(kegg_dbxref)
# remember the ID we just saw
last_qry_id = this_qry_id