def get_hmmdb_go_terms( acc, c ):
"""
This returns a list of bioannotation:GOAnnotation objects
"""
qry = "SELECT hg.go_id FROM hmm_go hg JOIN hmm ON hg.hmm_id=hmm.id WHERE hmm.accession = ?"
c.execute(qry, (acc,))
go_annots = list()
for row in c:
go = annotation.GOAnnotation(go_id=row[0], ev_code='ISM', with_from=acc)
go_annots.append(go)
return go_annots
def get_uniref_go_terms( acc, c ):
"""
This returns a list of bioannotation:GOAnnotation objects
"""
qry = """
SELECT us_go.go_id
FROM uniref_go us_go
WHERE us_go.id = ?
"""
c.execute(qry, (acc,))
go_annots = list()
for row in c:
go = annotation.GOAnnotation(go_id=row[0], ev_code='ISA', with_from=acc)
go_annots.append(go)
return go_annots
def get_uspdb_go_terms( acc, c ):
"""
This returns a list of bioannotation:GOAnnotation objects
"""
qry = """
SELECT us_go.go_id
FROM uniprot_sprot_go us_go
JOIN uniprot_sprot_acc us_acc ON us_go.id=us_acc.id
WHERE us_acc.accession = ?
"""
c.execute(qry, (acc,))
go_annots = list()
for row in c:
go = annotation.GOAnnotation(go_id=row[0], ev_code='ISA', with_from=acc)
go_annots.append(go)
return go_annots
def get_go_annotations(feat):
"""
Looks for sections like this to extract GO terms
CDS join(3366667..3366969,3463389..3463463)
/gene="ENSMUSG00000040653.6"
/protein_id="ENSMUSP00000149688.1"
/db_xref="EMBL:AC162384"
/db_xref="EMBL:CH466562"
/db_xref="GO:0005737"
/db_xref="GO:0042325"
"""
go_terms = list()
if 'db_xref' in feat.qualifiers:
for dbxref in feat.qualifiers['db_xref']:
m = re.match('GO:(.+)', dbxref)
if m:
go_terms.append(annotation.GOAnnotation(go_id=m.group(1)))
return go_terms
def parse_tmhmm_evidence(log_fh, polypeptides, htab_list):
'''
Reads a list of raw TMHMM evidence and a dict of polypeptides, adding annotation
attributes where possible.
Notes from the esteemed M Giglio:
The GO term to use would be GO:0016021 "integral component of membrane"
Or if you want to be more conservative you could go with GO:0016020 "membrane"
Depends on the evidence. For the prok pipe we are pretty conservative, we require five TMHMM
domains and then we call it putative integral membrane protein.
On ECO - in fact Marcus and I are the developers of ECO. It is an ontology of evidence types.
An annotation to an ECO term is used in conjunction with another annotation, like a GO term
(but many other types of annotation can, and are, used with ECO). It provides additional
information about the annotation. In fact for GO, the assignment of an evidence term along
with a GO term is a required part of a GO annotation. (ECO terms are the "evidence codes" in GO.)
INPUT: Expected TMHMM input (all HTML lines are skipped)
# CHARM010_V2.mRNA.887 Length: 904
# CHARM010_V2.mRNA.887 Number of predicted TMHs: 6
# CHARM010_V2.mRNA.887 Exp number of AAs in TMHs: 133.07638
# CHARM010_V2.mRNA.887 Exp number, first 60 AAs: 21.83212
# CHARM010_V2.mRNA.887 Total prob of N-in: 0.99994
# CHARM010_V2.mRNA.887 POSSIBLE N-term signal sequence
CHARM010_V2.mRNA.887 TMHMM2.0 inside 1 11
CHARM010_V2.mRNA.887 TMHMM2.0 TMhelix 12 34
CHARM010_V2.mRNA.887 TMHMM2.0 outside 35 712
CHARM010_V2.mRNA.887 TMHMM2.0 TMhelix 713 735
CHARM010_V2.mRNA.887 TMHMM2.0 inside 736 755
CHARM010_V2.mRNA.887 TMHMM2.0 TMhelix 756 773
CHARM010_V2.mRNA.887 TMHMM2.0 outside 774 782
CHARM010_V2.mRNA.887 TMHMM2.0 TMhelix 783 805
CHARM010_V2.mRNA.887 TMHMM2.0 inside 806 809
CHARM010_V2.mRNA.887 TMHMM2.0 TMhelix 810 832
CHARM010_V2.mRNA.887 TMHMM2.0 outside 833 871
CHARM010_V2.mRNA.887 TMHMM2.0 TMhelix 872 894
CHARM010_V2.mRNA.887 TMHMM2.0 inside 895 904
'''
# The number of helices spanning the membrane required before counted as a membrane protein
MIN_HELICAL_SPANS = 3
# For successful matches, this is the product name which gets applied
GENE_PRODUCT_NAME = 'Putative integral membrane protein'
for file in utils.read_list_file(htab_list):
last_qry_id = None
current_helix_count = 0
for line in open(file):
if line.startswith('<'): continue
m = re.match("# (.+?)\s+Length: \d+", line)
if m:
current_id = m.group(1)
# purge previous result
if current_helix_count >= MIN_HELICAL_SPANS:
annot = polypeptides[last_qry_id].annotation
if annot.product_name == DEFAULT_PRODUCT_NAME:
annot.product_name = GENE_PRODUCT_NAME
log_fh.write(
"INFO: {0}: Updated product name to '{1}' because it had {2} TMHelix domains predicted by TMHMM\n"
.format(last_qry_id, annot.product_name,
current_helix_count))
else:
log_fh.write(
"INFO: {0}: TMHMM predicted {1} TMHelix domains but gene product name unchanged because of previous assignment\n"
.format(last_qry_id, current_helix_count))
## we add the GO terms no matter what
annot.add_go_annotation(
annotation.GOAnnotation(go_id='0016021'))
# reset
last_qry_id = current_id
current_helix_count = 0
continue
cols = line.split()
if len(cols) == 5 and cols[2] == 'TMhelix':
current_helix_count += 1