def _fn(x, eid):
out = {'gene': eid, 'translation': []}
def mapping(transcript_id, protein_id):
trid = transcript_id and transcript_id != '\\N' and transcript_id or None
pid = protein_id and protein_id != '\\N' and protein_id or None
if trid and pid:
out['translation'].append({"rna": trid, "protein": pid})
if isinstance(x, list):
transcript_li = []
protein_li = []
for _x in x:
if _x[0] and _x[0] != '\\N':
transcript_li.append(_x[0])
if _x[1] and _x[1] != '\\N':
protein_li.append(_x[1])
mapping(_x[0], _x[1])
if transcript_li:
out['transcript'] = normalized_value(transcript_li)
if protein_li:
out['protein'] = normalized_value(protein_li)
else:
if x[0] and x[0] != '\\N':
out['transcript'] = x[0]
if x[1] and x[1] != '\\N':
out['protein'] = x[1]
mapping(x[0], x[1])
return out
def _ff(d):
(taxid, symbol, locus_tag, synonyms, dbxrefs, map_location,
description, type_of_gene, other_designations,
modification_date) = d
out = dict(taxid=int(taxid), symbol=symbol, name=description)
if map_location != '-':
out['map_location'] = map_location
if type_of_gene != '-':
out['type_of_gene'] = type_of_gene
if synonyms != '-':
out['alias'] = normalized_value(synonyms.split('|'))
if locus_tag != '-':
out['locus_tag'] = locus_tag
if other_designations != "-":
out['other_names'] = normalized_value(
other_designations.split('|'))
# when merged, this will become the default timestamp
# as of 2017/12/10, some timestamps can have different formats
if len(modification_date) > 8:
out["_timestamp"] = datetime.datetime.strptime(
modification_date, "%m/%d/%Y %H:%M:%S")
else:
out["_timestamp"] = datetime.datetime.strptime(
modification_date, "%Y%m%d")
for x in dbxrefs.split('|'):
if x == '-':
continue
xd = x.split(':')
if len(xd) == 3 and xd[0] == xd[1] and \
xd[0] in ['VGNC', 'HGNC', 'MGI']:
# a fix for NCBI bug for dup xref prefix, 'HGNC:HGNC:36328'
xd = xd[1:]
try:
_db, _id = xd
except:
print(repr(x))
raise
# we don't need ensembl xref from here, we will get it from
# Ensembl directly
if _db.lower() in ['ensembl', 'imgt/gene-db']:
# we don't need 'IMGT/GENE-DB" xref either, because they
# are mostly the same as gene symbol
continue
# add "MGI:" prefix for MGI ids.
if _db.lower() == 'mgi':
_id = "MGI:" + _id
out[_db] = _id
return out
def _ff(d):
out = {'rna': [], 'protein': [], 'genomic': [], 'translation': []}
for rna, prot, dna in d:
if rna == '-': rna = None
if prot == '-': prot = None
if dna == '-': dna = None
if rna is not None:
out['rna'].append(rna)
if prot is not None:
out['protein'].append(prot)
if dna is not None:
out['genomic'].append(dna)
if rna and prot:
out['translation'].append({'rna': rna, 'protein': prot})
# remove dup
for k in out:
out[k] = normalized_value(out[k])
# remove empty rna/protein/genomic field
_out = {}
for k, v in out.items():
if v:
_out[k] = v
if _out:
_out = {self.fieldname: _out}
return _out
def load(self, aslist=False):
if self.species_li:
_includefn = lambda ld: int(ld[0]) in self.taxid_set and ld[
1] != '-'
else:
_includefn = lambda ld: ld[1] != '-'
gene2retired = tab2dict(self.datafile, (1, 2),
0,
alwayslist=1,
includefn=_includefn)
gene2retired = dict_convert(
gene2retired,
valuefn=lambda x: normalized_value([int(xx) for xx in x]))
gene_d = {}
for gid, retired in gene2retired.items():
gene_d[gid] = {'retired': retired}
if aslist:
return dict_to_list(gene_d)
else:
return gene_d