def process_on_hold(on_hold, complexes_by_id, complexes):
on_hold_next = []
for name, components, id_ in on_hold:
components = [
[comp.components for comp in complexes_by_id[comp_id]]
if comp_id in complexes_by_id else ((comp_id, ), )
for comp_id in components
]
for components0 in itertools.product(*components):
this_components = list(itertools.chain(*components0))
if any(
comp.startswith('SIGNOR-C')
for comp in this_components):
on_hold_next.append((name, this_components, id_))
else:
cplex = intera.Complex(
name=name.replace('"', '').strip(),
components=this_components,
sources='SIGNOR',
ids=id_,
)
complexes[cplex.__str__()] = cplex
complexes_by_id[id_].add(cplex)
return on_hold_next, complexes_by_id, complexes
def get_partners(components, sources, references):
return {(comp[0] if len(comp) == 1 else intera.Complex(
components=comp,
sources=sources,
references=references,
))
for comp in components}
def name_method(rec):
comp = get_stoichiometry(rec)
cplex = intera.Complex(
name=rec['complex_name'],
components=comp,
sources='CellPhoneDB',
ids=rec['complex_name'],
)
return cplex
def _icellnet_get_entity(components, references):
if len(components) > 1:
return intera.Complex(
components=components,
sources='ICELLNET',
references=references,
)
elif len(components) == 1:
return components[0]
def _cellchatdb_process_complexes(raw, organism = 9606):
if isinstance(raw, dict):
raw = raw['complex']
ncbi_tax_id = _cellchatdb_organism(organism)
complexes = {}
for row in raw.itertuples():
genesymbols = [c for c in row[1:-1] if c]
uniprots = [
mapping.map_name(
gs,
'genesymbol',
'uniprot',
ncbi_tax_id = ncbi_tax_id,
)
for gs in genesymbols
]
uniprots = [up for up in uniprots if up]
for components in itertools.product(*uniprots):
cplex = intera.Complex(
name = row.rownames,
components = components,
sources = 'CellTalkDB',
ncbi_tax_id = ncbi_tax_id,
)
complexes[cplex.__str__()] = cplex
return complexes
def process_complex(ids, symbols, types):
if ids not in cplexes:
if not all(t == 'gene' for t in types):
cplexes[ids] = set()
uniprots = [
process_protein(id_, symbol)
for id_, symbol in zip(ids, symbols)
]
this_cplexes = {
intera.Complex(
components=components,
sources='KEGG-MEDICUS',
)
for components in itertools.product(*uniprots)
}
cplexes[ids] = this_cplexes
return cplexes[ids]
def signor_complexes(organism=9606):
#TODO: implement organism
def process_on_hold(on_hold, complexes_by_id, complexes):
on_hold_next = []
for name, components, id_ in on_hold:
components = [
[comp.components for comp in complexes_by_id[comp_id]]
if comp_id in complexes_by_id else ((comp_id, ), )
for comp_id in components
]
for components0 in itertools.product(*components):
this_components = list(itertools.chain(*components0))
if any(
comp.startswith('SIGNOR-C')
for comp in this_components):
on_hold_next.append((name, this_components, id_))
else:
cplex = intera.Complex(
name=name.replace('"', '').strip(),
components=this_components,
sources='SIGNOR',
ids=id_,
)
complexes[cplex.__str__()] = cplex
complexes_by_id[id_].add(cplex)
return on_hold_next, complexes_by_id, complexes
complexes = {}
on_hold = []
families = signor_protein_families(organism=organism)
url = urls.urls['signor']['complexes']
c = curl.Curl(
url,
binary_data=[(b'submit', b'Download complex data')],
large=True,
)
_ = next(c.result)
complexes_by_id = collections.defaultdict(set)
for rec in c.result:
rec = rec.split(';')
components = [u.strip('\n\r" ') for u in rec[2].split(',')]
components = [
families[comp] if comp in families else [comp]
for comp in components
]
for this_components in itertools.product(*components):
# some complex contains other complexes
if any(comp.startswith('SIGNOR-C') for comp in this_components):
on_hold.append((rec[1], this_components, rec[0]))
else:
cplex = intera.Complex(
name=rec[1].replace('"', '').strip(),
components=this_components,
sources='Signor',
ids=rec[0],
)
complexes[cplex.__str__()] = cplex
complexes_by_id[rec[0]].add(cplex)
while True:
# complexes are defined recursively
count_on_hold = len(on_hold)
on_hold, complexes_by_id, complexes = (process_on_hold(
on_hold, complexes_by_id, complexes))
if len(on_hold) == count_on_hold:
break
return complexes