def li2012_dmi():
"""
Converts table read by ``pypath.inputs.li2012.get_li2012`` to
list of ``pypath.internals.intera.DomainMotif`` objects.
Translates GeneSymbols to UniProt IDs.
"""
result = {}
nondigit = re.compile(r'[^\d]+')
se = uniprot_input.swissprot_seq(isoforms=True)
data = get_li2012()
for l in data:
subs_protein = l[1].split('/')[0]
tk_protein = l[2].split()[0]
reader_protein = l[3].split()[0]
subs_uniprots = mapping.map_name(
subs_protein,
'genesymbol',
'uniprot',
)
tk_uniprots = mapping.map_name(tk_protein, 'genesymbol', 'uniprot')
reader_uniprots = mapping.map_name(reader_protein, 'genesymbol',
'uniprot')
subs_resnum = int(non_digit.sub('', l[1].split('/')[1]))
for su in subs_uniprots:
if su in se:
subs_iso = None
for iso, s in iteritems(se[su].isof):
if se[su].get(subs_resnum, isoform=iso) == 'Y':
subs_iso = iso
break
if subs_iso:
start = min(1, subs_resnum - 7)
end = max(subs_resnum + 7, len(se[su].isof[subs_iso]))
for ku in tk_uniprots:
res = intera.Residue(subs_resnum,
'Y',
su,
isoform=subs_iso)
mot = intera.Motif(su,
start,
end,
isoform=subs_iso,
instance=se[su].get(
start, end, isoform=subs_iso))
ptm = intera.Ptm(su,
motif=mot,
residue=res,
isoform=subs_iso,
source='Li2012')
dom = intera.Domain(ku)
dommot = intera.DomainMotif(domain=dom,
ptm=ptm,
sources=['Li2012'])
result = {}
return result
def phosphosite_ptms(organism='human'):
"""
Downloads the phosphorylation site dataset from PhosphoSitePlus.
"""
result = []
url = urls.urls['psite_p']['url']
nondigit = re.compile(r'[^\d]+')
remot = re.compile(r'(_*)([A-Za-z]+)(_*)')
c = curl.Curl(url, silent=False, large=True)
data = c.result
for _ in xrange(4):
null = c.result.readline()
for r in data:
r = r.split('\t')
if len(r) > 9 and (organism is None or r[6] == organism):
uniprot = r[2]
isoform = 1 if '-' not in uniprot else int(uniprot.split('-')[1])
uniprot = uniprot.split('-')[0]
typ = r[3].lower()
if len(typ) == 0:
typ = r[4].split('-')[1] if '-' in r[4] else None
aa = r[4][0]
num = int(nondigit.sub('', r[4]))
motif = remot.match(r[9])
if motif:
start = num - 7 + len(motif.groups()[0])
end = num + 7 - len(motif.groups()[2])
instance = r[9].replace('_', '').upper()
else:
start = None
end = None
instance = None
res = intera.Residue(num, aa, uniprot, isoform=isoform)
mot = intera.Motif(uniprot,
start,
end,
instance=instance,
isoform=isoform)
ptm = intera.Ptm(uniprot,
typ=typ,
motif=mot,
residue=res,
source='PhosphoSite',
isoform=isoform)
result.append(ptm)
return result
def translate_ptm(self, ptm):
tptms = self.translate_site(
ptm.protein,
ptm.residue.name,
ptm.residue.number,
ptm.residue.isoform,
ptm.typ,
)
result = []
for x in tptms:
se = self.get_seq(x[0])
if (se is None or x[1] not in se.isof) and self.strict:
continue
res = intera.Residue(
number=x[3],
name=x[2],
protein=x[0],
isoform=x[1],
ncbi_tax_id=self.target,
)
start, end, region = (se.get_region(x[3], isoform=x[1])
if se is not None and x[1] in se.isof else
(None, None, None))
mot = intera.Motif(
protein=x[0],
start=start,
end=end,
instance=region,
isoform=x[1],
ncbi_tax_id=self.target,
)
ptm = intera.Ptm(
protein=x[0],
motif=mot,
residue=res,
typ=x[5],
isoform=x[1],
evidences=ptm.evidences,
ncbi_tax_id=self.target,
)
result.append(ptm)
return result
def _process(self, p):
# human leukocyte antigenes result a result an
# extremely high number of combinations
if (not p['kinase'] or (isinstance(p['substrate'], common.basestring)
and p['substrate'].startswith('HLA'))):
return
if not isinstance(p['kinase'], list):
p['kinase'] = [p['kinase']]
kinase_ups = mapping.map_names(
p['kinase'],
self.id_type_enzyme,
'uniprot',
ncbi_tax_id=self.ncbi_tax_id,
)
substrate_ups_all = set()
for sub_id_type in self.id_type_substrate:
if isinstance(sub_id_type, (list, tuple)):
sub_id_type, sub_id_attr = sub_id_type
else:
sub_id_attr = 'substrate'
substrate_ups_all.update(
set(
mapping.map_name(
p[sub_id_attr],
sub_id_type,
'uniprot',
self.ncbi_tax_id,
)))
# looking up sequences in all isoforms:
substrate_ups = []
for s in substrate_ups_all:
if 'substrate_isoform' in p and p['substrate_isoform']:
substrate_ups.append((s, p['substrate_isoform']))
else:
se = self.get_seq(s)
if se is None:
continue
for isof in se.isoforms():
if 'instance' in p and p['instance'] is not None:
if se.match(
p['instance'],
p['start'],
p['end'],
isoform=isof,
):
substrate_ups.append((s, isof))
else:
if se.match(
p['resaa'],
p['resnum'],
isoform=isof,
):
substrate_ups.append((s, isof))
if self.trace:
if p['substrate'] not in self.sub_ambig:
self.sub_ambig[p['substrate']] = substrate_ups
for k in p['kinase']:
if k not in self.kin_ambig:
self.kin_ambig[k] = kinase_ups
# generating report on non matching substrates
if len(substrate_ups) == 0:
for s in substrate_ups_all:
se = self.get_seq(s[0])
if se is None:
continue
self.nomatch.append((
s[0],
s[1],
(
p['substrate_refseq']
if 'substrate_refseq' in p else '',
s,
p['instance'],
se.get(p['start'], p['end']),
),
))
# building objects representing the enzyme-substrate interaction(s)
if 'typ' not in p:
p['typ'] = 'phosphorylation'
_resources = tuple(
(self.input_param.
get_via(name) if hasattr(self.input_param, 'get_via') else name)
for name in (p['databases'] if 'databases' in p else ()))
_resources += ((self.name, ) if isinstance(
self.input_param, common.basestring) else (self.input_param, ))
# collecting the evidences
evidences = evidence.Evidences(
evidence.Evidence(resource=_res,
references=p['references'] if 'references' in
p else None) for _res in _resources)
for s in substrate_ups:
# building the objects representing the substrate
se = self.get_seq(s[0])
if se is None:
continue
res = intera.Residue(
p['resnum'],
p['resaa'],
s[0],
isoform=s[1],
ncbi_tax_id=self.ncbi_tax_id,
)
if 'instance' not in p or p['instance'] is None:
reg = se.get_region(
p['resnum'],
p['start'] if 'start' in p else None,
p['end'] if 'end' in p else None,
isoform=s[1],
)
if reg is not None:
p['start'], p['end'], p['instance'] = reg
mot = intera.Motif(
s[0],
p['start'],
p['end'],
instance=p['instance'],
isoform=s[1],
ncbi_tax_id=self.ncbi_tax_id,
)
ptm = intera.Ptm(
s[0],
motif=mot,
residue=res,
typ=p['typ'],
evidences=evidences,
isoform=s[1],
ncbi_tax_id=self.ncbi_tax_id,
)
for k in kinase_ups:
if (not self.allow_mixed_organisms
and (self.get_taxon(k) != self.ncbi_tax_id
or self.get_taxon(s[0]) != self.ncbi_tax_id)):
continue
# the enzyme (kinase)
dom = intera.Domain(
protein=k,
ncbi_tax_id=self.ncbi_tax_id,
)
dommot = intera.DomainMotif(
domain=dom,
ptm=ptm,
evidences=evidences,
)
if hasattr(self.input_param, 'extra_attrs'):
for attr, key in iteritems(self.input_param.extra_attrs):
if key in p:
setattr(dommot, attr, p[key])
yield dommot
def phosphosite_enzyme_substrate(
raw=True,
organism='human',
strict=True,
):
"""
Downloads and preprocesses phosphorylation site data from PhosphoSitePlus.
"""
url = urls.urls['psite_kin']['url']
c = curl.Curl(
url,
silent=False,
compr='gz',
encoding='iso-8859-1',
large=True,
)
orto = {}
data = c.result
cols = {
'kinase': 2,
'kinase_org': 3,
'substrate': 6,
'substrate_org': 8,
'residue': 9,
'motif': 11
}
data = inputs_common.read_table(
cols=cols,
fileObject=data,
sep='\t',
hdr=4,
)
result = []
non_digit = re.compile(r'[^\d.-]+')
motre = re.compile(r'(_*)([A-Za-z]+)(_*)')
for r in data:
if organism is None or \
((r['kinase_org'] == organism or not strict) and \
r['substrate_org'] == organism):
if r['kinase_org'] != organism:
korg = r['kinase_org']
# attempting to map by orthology:
if korg in taxonomy.taxa and organism in taxonomy.taxa:
ktaxid = taxonomy.taxa[korg]
taxid = taxonomy.taxa[organism]
if korg not in orto:
orto[korg] = homology.homologene_dict(
ktaxid,
taxid,
'refseqp',
)
korg_refseq = mapping.map_name(r['kinase'], 'uniprot',
'refseqp', ktaxid)
kin_uniprot = \
list(
itertools.chain(
*map(
lambda ors:
mapping.map_name(ors,
'refseqp',
'uniprot',
taxid),
itertools.chain(
*map(
lambda rs:
orto[korg][rs],
filter(
lambda rs:
rs in orto[korg],
korg_refseq
)
)
)
)
)
)
else:
kin_uniprot = [r['kinase']]
for kinase in kin_uniprot:
r['resaa'] = r['residue'][0]
r['resnum'] = int(non_digit.sub('', r['residue'][1:]))
mot = motre.match(r['motif'])
# excluding e.g. Q12809_VAR_014388
r['substrate'] = r['substrate'].split('_')[0]
sisoform = 1 if '-' not in r['substrate'] else \
int(r['substrate'].split('-')[1])
r['substrate'] = r['substrate'].split('-')[0]
kisoform = (1 if '-' not in kinase else int(
kinase.split('-')[1]))
kinase = kinase.split('-')[0]
r['substrate'] = r['substrate'].split('-')[0]
if mot:
r['start'] = r['resnum'] - 7 + len(mot.groups()[0])
r['end'] = r['resnum'] + 7 - len(mot.groups()[2])
r['instance'] = r['motif'].replace('_', '').upper()
else:
r['start'] = None
r['end'] = None
r['instance'] = None
if raw:
r['kinase'] = kinase
result.append(r)
else:
res = intera.Residue(r['resnum'],
r['resaa'],
r['substrate'],
isoform=sisoform)
mot = intera.Motif(r['substrate'],
r['start'],
r['end'],
instance=r['instance'],
isoform=sisoform)
ptm = intera.Ptm(protein=r['substrate'],
residue=res,
motif=mot,
typ='phosphorylation',
source='PhosphoSite',
isoform=sisoform)
dom = intera.Domain(protein=kinase, isoform=kisoform)
dommot = intera.DomainMotif(domain=dom,
ptm=ptm,
sources=['PhosphoSite'])
result.append(dommot)
return result