def equiv(d, verbose):
if str(d) == 'entrez_info':
with open('entrez-gene-ids.beleq', 'w') as fp:
for gene_id in d.get_eq_values():
uid = uuid.uuid4()
fp.write(delim.join((gene_id, str(uid)))+'\n')
entrez_eq[gene_id] = uuid.uuid4()
make_eq_dict(d)
elif str(d) == 'hgnc':
with open('hgnc-approved-symbols.beleq', 'w') as fp:
for approved_symbol in d.get_eq_values():
if '~withdrawn' in approved_symbol:
continue
new_id = to_entrez('HGNC:'+approved_symbol)
if new_id is None:
# keep track of which hgnc genes need new uuids (dont map to entrez)
hgnc_list.append(approved_symbol)
# generate new uuid
uid = uuid.uuid4()
fp.write(delim.join((approved_symbol, str(uid)))+'\n')
hgnc_eq[approved_symbol] = uid
else:
# use the entrez uuid
uid = entrez_eq.get(new_id)
fp.write(delim.join((approved_symbol, str(uid)))+'\n')
hgnc_eq[approved_symbol] = uid
elif str(d) == 'mgi':
with open('mgi-approved-symbols.beleq', 'w') as fp:
for marker_symbol in d.get_eq_values():
new_id = to_entrez('MGI:'+marker_symbol)
if new_id is None:
# keep track of which genes need new uuids (dont map to entrez)
mgi_list.append(marker_symbol)
# generate new uuid
uid = uuid.uuid4()
fp.write(delim.join((marker_symbol, str(uid)))+'\n')
mgi_eq[marker_symbol] = uid
else:
# use the entrez uuid
uid = entrez_eq.get(new_id)
fp.write(delim.join((marker_symbol, str(uid)))+'\n')
mgi_eq[marker_symbol] = uid
elif str(d) == 'rgd':
with open('rgd-approved-symbols.beleq', 'w') as fp:
for symbol in d.get_eq_values():
new_id = to_entrez('RGD:'+symbol)
if new_id is None:
# keep track of which genes need new uuids (dont map to entrez)
rgd_list.append(symbol)
# generate new uuid
uid = uuid.uuid4()
fp.write(delim.join((symbol, str(uid)))+'\n')
rgd_eq[symbol] = uid
else:
# use the entrez uuid
uid = entrez_eq.get(new_id)
fp.write(delim.join((symbol, str(uid)))+'\n')
rgd_eq[symbol] = uid
elif str(d) == 'swiss':
with open('swissprot-entry-names.beleq', 'w') as fp:
# dbrefs is a dict, i.e { reference_type : id_of_that_gene}
for name, dbrefs, accessions in d.get_eq_values():
target_pool = ['HGNC', 'MGI', 'RGD']
gene_ids = []
alt_ids = []
if dbrefs is not None:
for k, v in dbrefs.items():
if k == 'GeneId':
gene_ids.extend(v)
if k in target_pool:
# could be MGI or RGD or HGNC ids
alt_ids.extend(v)
if len(gene_ids) == 1:
temp_id = entrez_eq.get(gene_ids[0])
if temp_id is None:
uid = uuid.uuid4()
fp.write(delim.join((name, str(uid)))+'\n')
sp_eq[name] = uid
else:
uid = entrez_eq.get(gene_ids[0])
fp.write(delim.join((name, str(uid)))+'\n')
sp_eq[name] = uid
elif len(gene_ids) == 0:
# are there hgnc, mgi or rgd refs?
if len(alt_ids) == 0:
uid = uuid.uuid4()
fp.write(delim.join((name, str(uid)))+'\n')
sp_eq[name] = uid
sp_list.append(name)
elif len(alt_ids) == 1:
a_id = alt_ids[0]
if 'HGNC' in a_id:
hgnc_key = namespaces.hgnc_map.get(a_id)
uid = hgnc_eq.get(hgnc_key)
# SwissProt may be referring to a since-removed gene.
if uid is None:
uid = uuid.uuid4()
fp.write(delim.join((name, str(uid)))+'\n')
sp_eq[name] = uid
else:
sp_eq[name] = uid
elif 'MGI' in a_id:
mgi_key = namespaces.mgi_map.get(a_id)
uid = mgi_eq.get(mgi_key)
# SwissProt may be referring to a since-removed gene.
if uid is None:
uid = uuid.uuid4()
fp.write(delim.join((name, str(uid)))+'\n')
sp_eq[name] = uid
else:
sp_eq[name] = uid
else:
rgd_key = namespaces.rgd_map.get(a_id)
uid = rgd_eq.get(rgd_key)
# SwissProt may be referring to a since-removed gene.
if uid is None:
uid = uuid.uuid4()
fp.write(delim.join((name, str(uid)))+'\n')
sp_eq[name] = uid
else:
fp.write(delim.join((name, str(uid)))+'\n')
sp_eq[name] = uid
# > 1 alt_id then generate a new uuid
else:
uid = uuid.uuid4()
fp.write(delim.join((name, str(uid)))+'\n')
sp_eq[name] = uid
# > 1 entrez id than generate a new uuid
else:
uid = uuid.uuid4()
fp.write(delim.join((name, str(uid)))+'\n')
sp_eq[name] = uid
# finally, generate .beleq for accession data also
build_acc_data(accessions, name)
swiss_accessions_eq()
elif str(d) == 'affy':
with open('affy-probeset-ids.beleq', 'w') as fp:
for probe_id, gene_id in d.get_eq_values():
if gene_id is not None and '---' not in gene_id:
# need the space before and after '///' because that is how it is parsed.
entrez_ids = gene_id.split(' /// ')
# for 1 entrez mapping, use the entez uuid
if len(entrez_ids) == 1:
status = entrez_eq.get(entrez_ids[0])
if status is None:
uid = uuid.uuid4()
fp.write(delim.join((probe_id, str(uid)))+'\n')
affy_eq[probe_id] = uid
else:
uid = status
fp.write(delim.join((probe_id, str(uid)))+'\n')
affy_eq[probe_id] = uid
# we have > 1 entrez mapping, resolve to one.
else:
adjacent_list = []
for entrez_gene in entrez_ids:
refstatus = refseq.get(entrez_gene)
adjacent_list.append(ref_status.get(refstatus))
# zipping yields a list of tuples like [('5307',0), ('104',2), ('3043',None)]
# i.e. [(entrez_id, refseq_status)]
list_of_tuples = list(zip(entrez_ids, adjacent_list))
# get rid of all 'None' tuples (No entrez mapping)
list_of_tuples = [tup for tup in list_of_tuples if tup[1] is not None]
# no mapping, generate new uuid
if len(list_of_tuples) == 0:
uid = uuid.uuid4()
fp.write(delim.join((probe_id, str(uid)))+'\n')
affy_eq[probe_id] = uid
# multiple entrez, resolve by refseq status
else:
# min tuple is highest refseq status (0 the best)
min_tuple = min(list_of_tuples, key=lambda x: x[1])
min_refseq = min_tuple[1]
lowest_tuples = []
for item in list_of_tuples:
if item[1] == min_refseq:
lowest_tuples.append(item)
# if mutiple genes with same refseq, resolve by lowest gene #
target_tuple = min(lowest_tuples)
uid = entrez_eq.get(target_tuple[0])
fp.write(delim.join((probe_id, str(uid)))+'\n')
affy_eq[probe_id] = uid
# no entrez mapping, create a new uuid
else:
uid = uuid.uuid4()
fp.write(delim.join((probe_id, str(uid)))+'\n')
affy_eq[probe_id] = uid
# equiv for alt ids and names relies on the equivalence for
# primary ids being completely generated.
elif str(d) == 'chebi':
with open('chebi-ids.beleq', 'w') as fp, open('chebi-names_eq.beleq', 'w') as f:
# like Entrez, new uuid for primary ids only the FIRST time.
for primary_id in d.get_primary_ids():
uid = uuid.uuid4()
fp.write(delim.join((primary_id, str(uid)))+'\n')
chebi_id_eq[primary_id] = uid
for alt_id in d.get_alt_ids():
if alt_id not in chebi_id_eq:
# get its primary equivalent and use its uuid
primary = d.alt_to_primary(alt_id)
uid = chebi_id_eq[primary]
fp.write(delim.join((alt_id, str(uid)))+'\n')
chebi_id_eq[alt_id] = uid
for name in d.get_names():
primary = d.name_to_primary(name)
uid = chebi_id_eq.get(primary)
f.write(delim.join((name, str(uid)))+'\n')
chebi_name_eq[name] = uid
elif str(d) == 'pubchem_equiv':
with open('pubchem_eq.beleq', 'w') as fp:
for sid, source, cid in d.get_eq_values():
if 'ChEBI' in source and cid is not None: # <-- verify that NO PubChem CID == 'None'
# use the CHEBI uuid
chebi_equiv = source.split(':')[1]
uid = chebi_id_eq.get(chebi_equiv)
fp.write(delim.join((sid, str(uid)))+'\n')
pub_eq_dict[sid] = uid
else:
# generate a new uuid
uid = uuid.uuid4()
fp.write(delim.join((sid, str(uid)))+'\n')
elif str(d) == 'gobp':
with open('go-biological-processes-names.beleq', 'w') as gobp, \
open('go-biological-processes-ids.beleq', 'w') as gobp_id:
for vals in d.get_eq_values():
termid, termname = vals
uid = uuid.uuid4()
gobp_id.write(delim.join((termid, str(uid)))+'\n')
gobp.write(delim.join((termname, str(uid)))+'\n')
gobp_eq_dict[termname] = uid
# GO is the baseline for processes, so new uuids the first time.
elif str(d) == 'gocc':
with open('go-cellular-component-terms.beleq', 'w') as gocc, \
open('go-cellular-component-ids.beleq', 'w') as gocc_id:
for vals in d.get_eq_values():
termid, termname = vals
uid = uuid.uuid4()
gocc_id.write(delim.join((termid, str(uid)))+'\n')
gocc.write(delim.join((termname, str(uid)))+'\n')
gocc_eq_dict[termid] = uid
elif str(d) == 'do':
# assign DO a new uuid and use as the primary for diseases
with open('disease-ontology-names.beleq', 'w') as dn, \
open('disease-ontology-ids.beleq', 'w') as di:
for vals in d.get_eq_values():
name, id = vals
uid = uuid.uuid4()
dn.write(delim.join((name, str(uid)))+'\n')
di.write(delim.join((id, str(uid)))+'\n')
do_eq_dict[name] = uid
elif str(d) == 'sdis_to_do':
# try to resolve sdis terms to DO. If there is not one,
# assign a new uuid.
count = 0
sdis = parsed.load_data('sdis')
with open('selventa-legacy-diseases.beleq', 'w') as dof:
for vals in sdis.get_eq_values():
uid = None
sdis_term = vals
if d.has_equivalence(sdis_term):
count = count + 1
do_term = d.get_equivalence(sdis_term)
if do_term in do_eq_dict:
uid = do_eq_dict[do_term]
else:
uid = do_eq_dict[do_term.lower()]
else:
uid = uuid.uuid4()
dof.write(delim.join((sdis_term, str(uid)))+'\n')
if verbose:
print('Able to resolve ' +str(count)+ ' legacy disease terms to DO.')
elif str(d) == 'schem_to_chebi':
# try to resolve schem terms to CHEBI. If there is not one,
# assign a new uuid.
count = 0
schem = parsed.load_data('schem')
with open('selventa-legacy-chemical-names.beleq', 'w') as schemf:
for vals in schem.get_eq_values():
uid = None
schem_term = vals
if d.has_equivalence(schem_term):
count = count + 1
chebi_term = d.get_equivalence(schem_term)
if chebi_term in chebi_name_eq:
uid = chebi_name_eq[chebi_term]
elif chebi_term.lower() in chebi_name_eq:
uid = chebi_name_eq[chebi_term.lower()]
else:
uid = uuid.uuid4()
schemf.write(delim.join((schem_term, str(uid)))+'\n')
if verbose:
print('Able to resolve ' +str(count)+ ' legacy chemical terms to CHEBI.')
elif str(d) == 'mesh':
with open('mesh-cellular-locations.beleq', 'w') as mesha, \
open('mesh-diseases.beleq', 'w') as meshc, \
open('mesh-biological-processes.beleq', 'w') as meshg:
do_data = parsed.load_data('do')
for vals in d.get_eq_values():
ui, mh, mns, synonyms = vals
if any('A11.284' in branch for branch in mns):
# get GO equiv if there is one
uid = None
go_id = mg_eq.get(mh)
# meshcs_to_gocc contains OBSOLETE GO terms at the moment.
# It is possible this lookup will return None, in that
# case generate a new uuid.
if go_id is not None:
uid = gocc_eq_dict.get(go_id)
# try to find out why lookups fail - maybe OBSOLETE?
if uid is None:
if verbose:
print('Lookup failed for: '+str(go_id))
uid = uuid.uuid4()
else:
uid = uuid.uuid4()
mesha.write(delim.join((mh, str(uid)))+'\n')
elif any('C' in branch for branch in mns):
# does UI exist as a Xref in DO?
xref = do_data.get_xrefs('MSH:'+ui)
if xref:
uid = do_eq_dict[xref]
else:
uid = uuid.uuid4()
meshc.write(delim.join((mh, str(uid)))+'\n')
elif any('G' in branch for branch in mns):
# synonyms for MeSH
uid = None
for syn in synonyms:
# root 'G' branch in GOBP
for name in gobp_eq_dict:
if syn.lower() == name.lower():
uid = gobp_eq_dict.get(name)
if uid is None:
uid = uuid.uuid4()
meshg.write(delim.join((mh, str(uid)))+'\n')
for root, dirs, filenames in os.walk(working_dir):
for f in filenames:
if f in baseline_data:
data_tuple = baseline_data.get(f)
parser = data_tuple[PARSER_TYPE]("datasets/" + f)
if verbose:
parser.is_verbose()
print("Running " + str(parser))
for x in parser.parse():
parsed.build_data(x, str(parser))
print("Phase II ran in " + str(((time.time() - interval_time) / 60)) + " minutes")
print("\n======= Phase III, building namespaces =======")
interval_time = time.time()
# load parsed data to build namespaces
ei = parsed.load_data("entrez_info")
eh = parsed.load_data("entrez_history")
hg = parsed.load_data("hgnc")
mg = parsed.load_data("mgi")
rg = parsed.load_data("rgd")
sp = parsed.load_data("swiss")
af = parsed.load_data("affy")
g2 = parsed.load_data("gene2acc")
chebi = parsed.load_data("chebi")
schem = parsed.load_data("schem")
schem_to_chebi = parsed.load_data("schem_to_chebi")
sdis = parsed.load_data("sdis")
sdis_to_do = parsed.load_data("sdis_to_do")
gobp = parsed.load_data("gobp")
gocc = parsed.load_data("gocc")
# pub_eq = parsed.load_data('pubchem_equiv')
def equiv(d, verbose):
if str(d) == "entrez_info":
with open("entrez-gene-ids.beleq", "w") as fp:
for gene_id in d.get_eq_values():
uid = uuid.uuid4()
fp.write(delim.join((gene_id, str(uid))) + "\n")
entrez_eq[gene_id] = uuid.uuid4()
make_eq_dict(d)
elif str(d) == "hgnc":
with open("hgnc-approved-symbols.beleq", "w") as fp:
for approved_symbol in d.get_eq_values():
if "~withdrawn" in approved_symbol:
continue
new_id = to_entrez("HGNC:" + approved_symbol)
if new_id is None:
# keep track of which hgnc genes need new uuids (dont map to entrez)
hgnc_list.append(approved_symbol)
# generate new uuid
uid = uuid.uuid4()
fp.write(delim.join((approved_symbol, str(uid))) + "\n")
hgnc_eq[approved_symbol] = uid
else:
# use the entrez uuid
uid = entrez_eq.get(new_id)
fp.write(delim.join((approved_symbol, str(uid))) + "\n")
hgnc_eq[approved_symbol] = uid
elif str(d) == "mgi":
with open("mgi-approved-symbols.beleq", "w") as fp:
for marker_symbol in d.get_eq_values():
new_id = to_entrez("MGI:" + marker_symbol)
if new_id is None:
# keep track of which genes need new uuids (dont map to entrez)
mgi_list.append(marker_symbol)
# generate new uuid
uid = uuid.uuid4()
fp.write(delim.join((marker_symbol, str(uid))) + "\n")
mgi_eq[marker_symbol] = uid
else:
# use the entrez uuid
uid = entrez_eq.get(new_id)
fp.write(delim.join((marker_symbol, str(uid))) + "\n")
mgi_eq[marker_symbol] = uid
elif str(d) == "rgd":
with open("rgd-approved-symbols.beleq", "w") as fp:
for symbol in d.get_eq_values():
new_id = to_entrez("RGD:" + symbol)
if new_id is None:
# keep track of which genes need new uuids (dont map to entrez)
rgd_list.append(symbol)
# generate new uuid
uid = uuid.uuid4()
fp.write(delim.join((symbol, str(uid))) + "\n")
rgd_eq[symbol] = uid
else:
# use the entrez uuid
uid = entrez_eq.get(new_id)
fp.write(delim.join((symbol, str(uid))) + "\n")
rgd_eq[symbol] = uid
elif str(d) == "swiss":
with open("swissprot-entry-names.beleq", "w") as fp:
# dbrefs is a dict, i.e { reference_type : id_of_that_gene}
for name, dbrefs, accessions in d.get_eq_values():
target_pool = ["HGNC", "MGI", "RGD"]
gene_ids = []
alt_ids = []
if dbrefs is not None:
for k, v in dbrefs.items():
if k == "GeneId":
gene_ids.extend(v)
if k in target_pool:
# could be MGI or RGD or HGNC ids
alt_ids.extend(v)
if len(gene_ids) == 1:
temp_id = entrez_eq.get(gene_ids[0])
if temp_id is None:
uid = uuid.uuid4()
fp.write(delim.join((name, str(uid))) + "\n")
sp_eq[name] = uid
else:
uid = entrez_eq.get(gene_ids[0])
fp.write(delim.join((name, str(uid))) + "\n")
sp_eq[name] = uid
elif len(gene_ids) == 0:
# are there hgnc, mgi or rgd refs?
if len(alt_ids) == 0:
uid = uuid.uuid4()
fp.write(delim.join((name, str(uid))) + "\n")
sp_eq[name] = uid
sp_list.append(name)
elif len(alt_ids) == 1:
a_id = alt_ids[0]
if "HGNC" in a_id:
hgnc_key = namespaces.hgnc_map.get(a_id)
uid = hgnc_eq.get(hgnc_key)
# SwissProt may be referring to a since-removed gene.
if uid is None:
uid = uuid.uuid4()
fp.write(delim.join((name, str(uid))) + "\n")
sp_eq[name] = uid
else:
sp_eq[name] = uid
elif "MGI" in a_id:
mgi_key = namespaces.mgi_map.get(a_id)
uid = mgi_eq.get(mgi_key)
# SwissProt may be referring to a since-removed gene.
if uid is None:
uid = uuid.uuid4()
fp.write(delim.join((name, str(uid))) + "\n")
sp_eq[name] = uid
else:
sp_eq[name] = uid
else:
rgd_key = namespaces.rgd_map.get(a_id)
uid = rgd_eq.get(rgd_key)
# SwissProt may be referring to a since-removed gene.
if uid is None:
uid = uuid.uuid4()
fp.write(delim.join((name, str(uid))) + "\n")
sp_eq[name] = uid
else:
fp.write(delim.join((name, str(uid))) + "\n")
sp_eq[name] = uid
# > 1 alt_id then generate a new uuid
else:
uid = uuid.uuid4()
fp.write(delim.join((name, str(uid))) + "\n")
sp_eq[name] = uid
# > 1 entrez id than generate a new uuid
else:
uid = uuid.uuid4()
fp.write(delim.join((name, str(uid))) + "\n")
sp_eq[name] = uid
# finally, generate .beleq for accession data also
build_acc_data(accessions, name)
swiss_accessions_eq()
elif str(d) == "affy":
with open("affy-probeset-ids.beleq", "w") as fp:
for probe_id, gene_id in d.get_eq_values():
if gene_id is not None and "---" not in gene_id:
# need the space before and after '///' because that is how it is parsed.
entrez_ids = gene_id.split(" /// ")
# for 1 entrez mapping, use the entez uuid
if len(entrez_ids) == 1:
status = entrez_eq.get(entrez_ids[0])
if status is None:
uid = uuid.uuid4()
fp.write(delim.join((probe_id, str(uid))) + "\n")
affy_eq[probe_id] = uid
else:
uid = status
fp.write(delim.join((probe_id, str(uid))) + "\n")
affy_eq[probe_id] = uid
# we have > 1 entrez mapping, resolve to one.
else:
adjacent_list = []
for entrez_gene in entrez_ids:
refstatus = refseq.get(entrez_gene)
adjacent_list.append(ref_status.get(refstatus))
# zipping yields a list of tuples like [('5307',0), ('104',2), ('3043',None)]
# i.e. [(entrez_id, refseq_status)]
list_of_tuples = list(zip(entrez_ids, adjacent_list))
# get rid of all 'None' tuples (No entrez mapping)
list_of_tuples = [tup for tup in list_of_tuples if tup[1] is not None]
# no mapping, generate new uuid
if len(list_of_tuples) == 0:
uid = uuid.uuid4()
fp.write(delim.join((probe_id, str(uid))) + "\n")
affy_eq[probe_id] = uid
# multiple entrez, resolve by refseq status
else:
# min tuple is highest refseq status (0 the best)
min_tuple = min(list_of_tuples, key=lambda x: x[1])
min_refseq = min_tuple[1]
lowest_tuples = []
for item in list_of_tuples:
if item[1] == min_refseq:
lowest_tuples.append(item)
# if mutiple genes with same refseq, resolve by lowest gene #
target_tuple = min(lowest_tuples)
uid = entrez_eq.get(target_tuple[0])
fp.write(delim.join((probe_id, str(uid))) + "\n")
affy_eq[probe_id] = uid
# no entrez mapping, create a new uuid
else:
uid = uuid.uuid4()
fp.write(delim.join((probe_id, str(uid))) + "\n")
affy_eq[probe_id] = uid
# equiv for alt ids and names relies on the equivalence for
# primary ids being completely generated.
elif str(d) == "chebi":
with open("chebi-ids.beleq", "w") as fp, open("chebi-names_eq.beleq", "w") as f:
# like Entrez, new uuid for primary ids only the FIRST time.
for primary_id in d.get_primary_ids():
uid = uuid.uuid4()
fp.write(delim.join((primary_id, str(uid))) + "\n")
chebi_id_eq[primary_id] = uid
for alt_id in d.get_alt_ids():
if alt_id not in chebi_id_eq:
# get its primary equivalent and use its uuid
primary = d.alt_to_primary(alt_id)
uid = chebi_id_eq[primary]
fp.write(delim.join((alt_id, str(uid))) + "\n")
chebi_id_eq[alt_id] = uid
for name in d.get_names():
primary = d.name_to_primary(name)
uid = chebi_id_eq.get(primary)
f.write(delim.join((name, str(uid))) + "\n")
chebi_name_eq[name] = uid
elif str(d) == "pubchem_equiv":
with open("pubchem_eq.beleq", "w") as fp:
for sid, source, cid in d.get_eq_values():
if "ChEBI" in source and cid is not None: # <-- verify that NO PubChem CID == 'None'
# use the CHEBI uuid
chebi_equiv = source.split(":")[1]
uid = chebi_id_eq.get(chebi_equiv)
fp.write(delim.join((sid, str(uid))) + "\n")
pub_eq_dict[sid] = uid
else:
# generate a new uuid
uid = uuid.uuid4()
fp.write(delim.join((sid, str(uid))) + "\n")
elif str(d) == "gobp":
with open("go-biological-processes-names.beleq", "w") as gobp, open(
"go-biological-processes-ids.beleq", "w"
) as gobp_id:
for vals in d.get_eq_values():
termid, termname = vals
uid = uuid.uuid4()
gobp_id.write(delim.join((termid, str(uid))) + "\n")
gobp.write(delim.join((termname, str(uid))) + "\n")
gobp_eq_dict[termname] = uid
# GO is the baseline for processes, so new uuids the first time.
elif str(d) == "gocc":
with open("go-cellular-component-terms.beleq", "w") as gocc, open(
"go-cellular-component-ids.beleq", "w"
) as gocc_id:
for vals in d.get_eq_values():
termid, termname = vals
uid = uuid.uuid4()
gocc_id.write(delim.join((termid, str(uid))) + "\n")
gocc.write(delim.join((termname, str(uid))) + "\n")
gocc_eq_dict[termid] = uid
elif str(d) == "do":
# assign DO a new uuid and use as the primary for diseases
with open("disease-ontology-names.beleq", "w") as dn, open("disease-ontology-ids.beleq", "w") as di:
for vals in d.get_eq_values():
name, id = vals
uid = uuid.uuid4()
dn.write(delim.join((name, str(uid))) + "\n")
di.write(delim.join((id, str(uid))) + "\n")
do_eq_dict[name] = uid
elif str(d) == "sdis_to_do":
# try to resolve sdis terms to DO. If there is not one,
# assign a new uuid.
count = 0
sdis = parsed.load_data("sdis")
with open("selventa-legacy-diseases.beleq", "w") as dof:
for vals in sdis.get_eq_values():
uid = None
sdis_term = vals
if d.has_equivalence(sdis_term):
count = count + 1
do_term = d.get_equivalence(sdis_term)
if do_term in do_eq_dict:
uid = do_eq_dict[do_term]
else:
uid = do_eq_dict[do_term.lower()]
else:
uid = uuid.uuid4()
dof.write(delim.join((sdis_term, str(uid))) + "\n")
if verbose:
print("Able to resolve " + str(count) + " legacy disease terms to DO.")
elif str(d) == "schem_to_chebi":
# try to resolve schem terms to CHEBI. If there is not one,
# assign a new uuid.
count = 0
schem = parsed.load_data("schem")
with open("selventa-legacy-chemical-names.beleq", "w") as schemf:
for vals in schem.get_eq_values():
uid = None
schem_term = vals
if d.has_equivalence(schem_term):
count = count + 1
chebi_term = d.get_equivalence(schem_term)
if chebi_term in chebi_name_eq:
uid = chebi_name_eq[chebi_term]
elif chebi_term.lower() in chebi_name_eq:
uid = chebi_name_eq[chebi_term.lower()]
else:
uid = uuid.uuid4()
schemf.write(delim.join((schem_term, str(uid))) + "\n")
if verbose:
print("Able to resolve " + str(count) + " legacy chemical terms to CHEBI.")
elif str(d) == "mesh":
with open("mesh-cellular-locations.beleq", "w") as mesha, open("mesh-diseases.beleq", "w") as meshc, open(
"mesh-biological-processes.beleq", "w"
) as meshg:
do_data = parsed.load_data("do")
for vals in d.get_eq_values():
ui, mh, mns, synonyms = vals
if any("A11.284" in branch for branch in mns):
# get GO equiv if there is one
uid = None
go_id = mg_eq.get(mh)
# meshcs_to_gocc contains OBSOLETE GO terms at the moment.
# It is possible this lookup will return None, in that
# case generate a new uuid.
if go_id is not None:
uid = gocc_eq_dict.get(go_id)
# try to find out why lookups fail - maybe OBSOLETE?
if uid is None:
if verbose:
print("Lookup failed for: " + str(go_id))
uid = uuid.uuid4()
else:
uid = uuid.uuid4()
mesha.write(delim.join((mh, str(uid))) + "\n")
elif any("C" in branch for branch in mns):
# does UI exist as a Xref in DO?
xref = do_data.get_xrefs("MSH:" + ui)
if xref:
uid = do_eq_dict[xref]
else:
uid = uuid.uuid4()
meshc.write(delim.join((mh, str(uid))) + "\n")
elif any("G" in branch for branch in mns):
# synonyms for MeSH
uid = None
for syn in synonyms:
# root 'G' branch in GOBP
for name in gobp_eq_dict:
if syn.lower() == name.lower():
uid = gobp_eq_dict.get(name)
if uid is None:
uid = uuid.uuid4()
meshg.write(delim.join((mh, str(uid))) + "\n")
def make_namespace(d, verbose):
# build and write out the namespace values for each dataset
delim = '|'
if str(d) == 'entrez_info':
with open('entrez-gene-ids.belns', 'w') as fp:
# tuple of (gene_id, gene_type, description)
for vals in d.get_ns_values():
gene_id, gene_type, description = vals
if gene_type == 'miscRNA':
if 'microRNA' in description:
fp.write(delim.join((gene_id, 'GRM')) + '\n')
entrez_ns.add(gene_id)
else:
fp.write(delim.join((gene_id, 'GR')) + '\n')
entrez_ns.add(gene_id)
else:
fp.write(
delim.join((gene_id, entrez_encoding[gene_type])) +
'\n')
entrez_ns.add(gene_id)
elif str(d) == 'hgnc':
with open('hgnc-approved-symbols.belns', 'w') as fp:
for vals in d.get_ns_values():
approved_symb, locus_type, hgnc_id = vals
# withdrawn genes NOT included in this namespace
if locus_type is not 'withdrawn' and 'withdrawn' not in approved_symb:
fp.write(
delim.join((approved_symb,
hgnc_encoding[locus_type])) + '\n')
hgnc_ns.add(approved_symb)
hgnc_map[hgnc_id] = approved_symb
elif str(d) == 'mgi':
with open('mgi-approved-symbols.belns', 'w') as fp:
for vals in d.get_ns_values():
marker_symbol, feature_type, acc_id, marker_type = vals
if marker_type == 'Gene' or marker_type == 'Pseudogene':
fp.write(
delim.join((marker_symbol,
mgi_encoding[feature_type])) + '\n')
mgi_ns.add(marker_symbol)
mgi_map[acc_id] = marker_symbol
# withdrawn genes are NOT included in this namespace
elif str(d) == 'rgd':
with open('rgd-approved-symbols.belns', 'w') as fp:
for vals in d.get_ns_values():
symbol, gene_type, name, rgd_id = vals
if gene_type == 'miscrna' and 'microRNA' in name:
fp.write(delim.join((symbol, 'GRM')) + '\n')
rgd_ns.add(symbol)
elif gene_type == 'miscrna' and 'microRNA' not in name:
fp.write(delim.join((symbol, 'GR')) + '\n')
rgd_ns.add(symbol)
else:
if gene_type is not '':
fp.write(
delim.join((symbol, rgd_encoding[gene_type])) +
'\n')
rgd_ns.add(symbol)
rgd_map[rgd_id] = symbol
elif str(d) == 'swiss':
with open('swissprot-entry-names.belns', 'w') as fp, \
open('swissprot-accession-numbers.belns', 'w') as f:
for vals in d.get_ns_values():
gene_name, accessions = vals
fp.write(delim.join((gene_name, 'GRP')) + '\n')
sp_ns.add(gene_name)
for acc in accessions:
f.write(delim.join((acc, 'GRP')) + '\n')
sp_acc_ns.add(acc)
elif str(d) == 'affy':
with open('affy-probeset-ids.belns', 'w') as fp:
for vals in d.get_ns_values():
pid = vals
fp.write(delim.join((pid, 'R')) + '\n')
# if pid not in affy_ns_dict:
# affy_ns_dict[pid] = 'R'
elif str(d) == 'chebi':
with open('chebi-names.belns', 'w') as fp, \
open('chebi-ids.belns', 'w') as f:
for vals in d.get_ns_values():
name, primary_id, altIds = vals
fp.write(delim.join((name, 'A')) + '\n')
chebi_name_ns.add(name)
f.write(delim.join((primary_id, 'A')) + '\n')
chebi_id_ns.add(name)
if altIds:
for i in altIds:
if i not in chebi_id_ns:
f.write(delim.join((i, 'A')) + '\n')
chebi_id_ns.add(i)
elif str(d) == 'pubchem_namespace':
with open('pubchem-ids.belns', 'w') as fp:
for vals in d.get_ns_values():
pid = vals
fp.write(delim.join((pid, 'A')) + '\n')
pub_ns.add(pid)
elif str(d) == 'gobp':
with open('go-biological-processes-names.belns', 'w') as gobp, \
open('go-biological-processes-accession-numbers.belns', 'w') \
as gobp_id:
for vals in d.get_ns_values():
termid, termname, altids = vals
gobp.write(delim.join((termname, 'B')) + '\n')
gobp_id.write(delim.join((termid, 'B')) + '\n')
if altids is not None:
for alt in altids:
gobp_id.write(delim.join((alt, 'B')) + '\n')
elif str(d) == 'gocc':
with open('go-cellular-component-terms.belns', 'w') as gocc, \
open('go-cellular-component-accession-numbers.belns', 'w') \
as gocc_id:
for vals in d.get_ns_values():
termid, termname, altids, complex = vals
if complex:
gocc.write(delim.join((termname, 'C')) + '\n')
gocc_id.write(delim.join((termid, 'C')) + '\n')
for alt in altids:
gocc_id.write(delim.join((alt, 'C')) + '\n')
else:
gocc.write(delim.join((termname, 'A')) + '\n')
gocc_id.write(delim.join((termid, 'A')) + '\n')
if altids is not None:
for alt in altids:
gocc_id.write(delim.join((alt, 'A')) + '\n')
elif str(d) == 'schem':
schem_to_chebi = parsed.load_data('schem_to_chebi')
count = 0
with open('selventa-legacy-chemical-names.belns', 'w') as f:
for entry in d.get_ns_values():
# try to get a chebi equivalent, if there is one do not
# write this value to the new namespace
if schem_to_chebi.has_equivalence(entry):
count = count + 1
continue
else:
f.write(delim.join((entry, 'A')) + '\n')
if verbose:
print('Able to resolve ' + str(count) + ' SCHEM names to CHEBI.')
elif str(d) == 'sdis':
sdis_to_do = parsed.load_data('sdis_to_do')
count = 0
with open('selventa-legacy-diseases.belns', 'w') as f:
for entry in d.get_ns_values():
# try to get a do equivalent, if there is one do not
# write this value to the new namespace
if sdis_to_do.has_equivalence(entry):
count = count + 1
continue
else:
f.write(delim.join((entry, 'O')) + '\n')
if verbose:
print('Able to resolve ' + str(count) + ' SDIS names to DO.')
elif str(d) == 'mesh':
with open('mesh-cellular-locations.belns', 'w') as meshf, \
open('mesh-diseases.belns', 'w') as meshd, \
open('mesh-biological-processes.belns', 'w') as meshb:
for vals in d.get_ns_values():
ui, mh, mns, sts = vals
# all entries from A11.284 branch (abundances)
if any('A11.284' in branch for branch in mns):
meshf.write(delim.join((mh, 'A')) + '\n')
# all entries from the C branch (pathology)
elif any('C' in branch for branch in mns):
meshd.write(delim.join((mh, 'O')) + '\n')
# G branch (bio process) - exclude G01 G02 G15 G17 branches
elif any('G' in branch for branch in mns):
excluded = False
for branch in mns:
if branch.startswith('MN = G01') \
or branch.startswith('MN = G02') \
or branch.startswith('MN = G15') \
or branch.startswith('MN = G17'):
excluded = True
if not excluded:
meshb.write(delim.join((mh, 'B')) + '\n')
elif str(d) == 'do':
with open('disease-ontology-names.belns', 'w') as dn, \
open('disease-ontology-ids.belns', 'w') as di:
for vals in d.get_ns_values():
name, id = vals
dn.write(delim.join((name, 'O')) + '\n')
di.write(delim.join((id, 'O')) + '\n')
def make_namespace(d, verbose):
# build and write out the namespace values for each dataset
delim = '|'
if str(d) == 'entrez_info':
with open('entrez-gene-ids.belns', 'w') as fp:
# tuple of (gene_id, gene_type, description)
for vals in d.get_ns_values():
gene_id, gene_type, description = vals
if gene_type == 'miscRNA':
if 'microRNA' in description:
fp.write(delim.join((gene_id, 'GRM'))+'\n')
entrez_ns.add(gene_id)
else:
fp.write(delim.join((gene_id, 'GR'))+'\n')
entrez_ns.add(gene_id)
else:
fp.write(delim.join((gene_id, entrez_encoding[gene_type]))+'\n')
entrez_ns.add(gene_id)
elif str(d) == 'hgnc':
with open('hgnc-approved-symbols.belns', 'w') as fp:
for vals in d.get_ns_values():
approved_symb, locus_type, hgnc_id = vals
# withdrawn genes NOT included in this namespace
if locus_type is not 'withdrawn' and 'withdrawn' not in approved_symb:
fp.write(delim.join((approved_symb, hgnc_encoding[locus_type]))+'\n')
hgnc_ns.add(approved_symb)
hgnc_map[hgnc_id] = approved_symb
elif str(d) == 'mgi':
with open('mgi-approved-symbols.belns', 'w') as fp:
for vals in d.get_ns_values():
marker_symbol, feature_type, acc_id, marker_type = vals
if marker_type == 'Gene' or marker_type == 'Pseudogene':
fp.write(delim.join((marker_symbol, mgi_encoding[feature_type]))+'\n')
mgi_ns.add(marker_symbol)
mgi_map[acc_id] = marker_symbol
# withdrawn genes are NOT included in this namespace
elif str(d) == 'rgd':
with open('rgd-approved-symbols.belns', 'w') as fp:
for vals in d.get_ns_values():
symbol, gene_type, name, rgd_id = vals
if gene_type == 'miscrna' and 'microRNA' in name:
fp.write(delim.join((symbol, 'GRM'))+'\n')
rgd_ns.add(symbol)
elif gene_type == 'miscrna' and 'microRNA' not in name:
fp.write(delim.join((symbol, 'GR'))+'\n')
rgd_ns.add(symbol)
else:
if gene_type is not '':
fp.write(delim.join((symbol, rgd_encoding[gene_type]))+'\n')
rgd_ns.add(symbol)
rgd_map[rgd_id] = symbol
elif str(d) == 'swiss':
with open('swissprot-entry-names.belns', 'w') as fp, \
open('swissprot-accession-numbers.belns', 'w') as f:
for vals in d.get_ns_values():
gene_name, accessions = vals
fp.write(delim.join((gene_name, 'GRP'))+'\n')
sp_ns.add(gene_name)
for acc in accessions:
f.write(delim.join((acc, 'GRP'))+'\n')
sp_acc_ns.add(acc)
elif str(d) == 'affy':
with open('affy-probeset-ids.belns', 'w') as fp:
for vals in d.get_ns_values():
pid = vals
fp.write(delim.join((pid, 'R'))+'\n')
# if pid not in affy_ns_dict:
# affy_ns_dict[pid] = 'R'
elif str(d) == 'chebi':
with open('chebi-names.belns', 'w') as fp, \
open('chebi-ids.belns', 'w') as f:
for vals in d.get_ns_values():
name, primary_id, altIds = vals
fp.write(delim.join((name, 'A'))+'\n')
chebi_name_ns.add(name)
f.write(delim.join((primary_id, 'A'))+'\n')
chebi_id_ns.add(name)
if altIds:
for i in altIds:
if i not in chebi_id_ns:
f.write(delim.join((i, 'A'))+'\n')
chebi_id_ns.add(i)
elif str(d) == 'pubchem_namespace':
with open('pubchem-ids.belns', 'w') as fp:
for vals in d.get_ns_values():
pid = vals
fp.write(delim.join((pid, 'A'))+'\n')
pub_ns.add(pid)
elif str(d) == 'gobp':
with open('go-biological-processes-names.belns', 'w') as gobp, \
open('go-biological-processes-accession-numbers.belns', 'w') \
as gobp_id:
for vals in d.get_ns_values():
termid, termname, altids = vals
gobp.write(delim.join((termname, 'B'))+'\n')
gobp_id.write(delim.join((termid, 'B'))+'\n')
if altids is not None:
for alt in altids:
gobp_id.write(delim.join((alt, 'B'))+'\n')
elif str(d) == 'gocc':
with open('go-cellular-component-terms.belns', 'w') as gocc, \
open('go-cellular-component-accession-numbers.belns', 'w') \
as gocc_id:
for vals in d.get_ns_values():
termid, termname, altids, complex = vals
if complex:
gocc.write(delim.join((termname, 'C'))+'\n')
gocc_id.write(delim.join((termid, 'C'))+'\n')
for alt in altids:
gocc_id.write(delim.join((alt, 'C'))+'\n')
else:
gocc.write(delim.join((termname, 'A'))+'\n')
gocc_id.write(delim.join((termid, 'A'))+'\n')
if altids is not None:
for alt in altids:
gocc_id.write(delim.join((alt, 'A'))+'\n')
elif str(d) == 'schem':
schem_to_chebi = parsed.load_data('schem_to_chebi')
count = 0
with open('selventa-legacy-chemical-names.belns', 'w') as f:
for entry in d.get_ns_values():
# try to get a chebi equivalent, if there is one do not
# write this value to the new namespace
if schem_to_chebi.has_equivalence(entry):
count = count + 1
continue
else:
f.write(delim.join((entry, 'A'))+'\n')
if verbose:
print('Able to resolve ' +str(count)+ ' SCHEM names to CHEBI.')
elif str(d) == 'sdis':
sdis_to_do = parsed.load_data('sdis_to_do')
count = 0
with open('selventa-legacy-diseases.belns', 'w') as f:
for entry in d.get_ns_values():
# try to get a do equivalent, if there is one do not
# write this value to the new namespace
if sdis_to_do.has_equivalence(entry):
count = count + 1
continue
else:
f.write(delim.join((entry, 'O'))+'\n')
if verbose:
print('Able to resolve ' +str(count)+ ' SDIS names to DO.')
elif str(d) == 'mesh':
with open('mesh-cellular-locations.belns', 'w') as meshf, \
open('mesh-diseases.belns', 'w') as meshd, \
open('mesh-biological-processes.belns', 'w') as meshb:
for vals in d.get_ns_values():
ui, mh, mns, sts = vals
# all entries from A11.284 branch (abundances)
if any('A11.284' in branch for branch in mns):
meshf.write(delim.join((mh, 'A'))+'\n')
# all entries from the C branch (pathology)
elif any('C' in branch for branch in mns):
meshd.write(delim.join((mh, 'O'))+'\n')
# G branch (bio process) - exclude G01 G02 G15 G17 branches
elif any('G' in branch for branch in mns):
excluded = False
for branch in mns:
if branch.startswith('MN = G01') \
or branch.startswith('MN = G02') \
or branch.startswith('MN = G15') \
or branch.startswith('MN = G17'):
excluded = True
if not excluded:
meshb.write(delim.join((mh, 'B'))+'\n')
elif str(d) == 'do':
with open('disease-ontology-names.belns', 'w') as dn, \
open('disease-ontology-ids.belns', 'w') as di:
for vals in d.get_ns_values():
name, id = vals
dn.write(delim.join((name, 'O'))+'\n')
di.write(delim.join((id, 'O'))+'\n')