def __init__(self):
'''
Constructor
'''
# VARIABLES
self.umls_hpo2umls_d = {}
self.umls_mesh2umls_d = {}
self.inferred_hpo2mesh_d = {}
self.inferred_mesh2hpo_d = {}
self.hp2name_d = {}
self.mesh2name_d = {}
# ALGORITHM
# Create hpo2umls dict from UMLS mappings
for line in open(
'/home/nuria/workspace/repurposing-hetio/rephetio-dhimmelstein/hetionet+hpo/data/umls-hpo-2016aa.tsv',
'r').readlines():
if line.startswith('CUI'):
continue
line_l = line.strip('\n').split('\t')
cui = line_l[0]
hp_code = line_l[2]
hp_term = line_l[3]
util.add_elem_with_dictionary(self.umls_hpo2umls_d, hp_code, cui)
self.hp2name_d[hp_code] = hp_term
#print('{}\t{}\t{}'.format(cui,hp_code,hp_term))
# Create mesh2umls dict from UMLS mappings
for line in open(
'/home/nuria/workspace/repurposing-hetio/rephetio-dhimmelstein/hetionet+hpo/data/umls-mesh-2016aa.tsv',
'r').readlines():
if line.startswith('CUI'):
continue
line_l = line.strip('\n').split('\t')
cui = line_l[0]
mesh_code = line_l[2]
mesh_term = line_l[3]
util.add_elem_with_dictionary(self.umls_mesh2umls_d, mesh_code,
cui)
self.mesh2name_d[mesh_code] = mesh_term
#print('{}\t{}\t{}'.format(cui,mesh_code,mesh_term))
# Infer and create hpo2mesh dict
for hp in self.umls_hpo2umls_d:
hpCuis_s = set(self.umls_hpo2umls_d[hp])
for mesh in self.umls_mesh2umls_d:
meshCuis_s = set(self.umls_mesh2umls_d[mesh])
if len(hpCuis_s) > len(meshCuis_s) or len(hpCuis_s) == len(
meshCuis_s):
resta = hpCuis_s - meshCuis_s
else:
resta = meshCuis_s - hpCuis_s
if len(resta) == 0:
util.add_elem_with_dictionary(self.inferred_hpo2mesh_d, hp,
mesh)
util.add_elem_with_dictionary(self.inferred_mesh2hpo_d,
mesh, hp)
def map_do_orpha2do(path):
'''
Function to map disease identifiers from Orphanumber to DOID using DO extracted mappings
:param file: /home/nuria/workspace/repurposing-hetio/rephetio-dhimmelstein/hetionet+hpo/data/orphanet-disease-symptom.tsv
:return file: /home/nuria/workspace/repurposing-hetio/rephetio-dhimmelstein/hetionet+hpo/data/do-orpha2do-mappings.tsv
:return file: /home/nuria/workspace/repurposing-hetio/rephetio-dhimmelstein/hetionet+hpo/data/orphanet-do-doid_orphanum-hp.tsv
:return file: /home/nuria/workspace/repurposing-hetio/rephetio-dhimmelstein/hetionet+hpo/data/orphanet-diseases-do-orpha2do.tsv
'''
# IN
orphaDisPhe_p = '/home/nuria/workspace/repurposing-hetio/rephetio-dhimmelstein/hetionet+hpo/data/orphanet-disease-symptom.tsv'
with open('{}'.format(orphaDisPhe_p), 'r') as orphaDisPhe_f:
orphaDisPhe_l = orphaDisPhe_f.readlines()
orphaDisPhe_f.close()
# OUT
rd_sym_f = open("{}orphanet-do-doid_orphanum-hp.tsv".format(path), 'w')
rd_sym_f.write(
'orphanet_code\tdo_do:orphanet_cardinality\tdo_do_mapping\torphanet_term\thp_code\thp_term\n'
)
rd_mappings_f = open("{}do-orpha2do-mappings.tsv".format(path, ), 'w')
rd_mappings_f.write("orphanet\tdo\n")
orpha_mappings_f = open(
"{}orphanet-diseases-do-orpha2do.tsv".format(path, ), 'w')
orpha_mappings_f.write('orphanumber\tdoid\n')
# VARIABLES
orpha2do_dct = {}
do_orpha_dct = {}
orpha_orpha_dct = {}
# ALGORITHM
# import do doid2orpha mappings dictionary
do_owl = "/home/nuria/workspace/repurposing-hetio/rephetio-dhimmelstein/hetionet+hpo/data/doid.owl"
do_term = do.term(do_owl)
do_do2orpha_dct = do_term.do2orpha
# define reverse mappings, orpha to do
for doid, orphanumber_l in do_do2orpha_dct.items():
for orpha in orphanumber_l:
util.add_elem_with_dictionary(orpha2do_dct, orpha, doid)
do_orpha_dct[orpha] = 1
# write down the do mappings orpha2do
for orpha in orpha2do_dct:
for key in orpha2do_dct[orpha].keys():
rd_mappings_f.write("{}\t{}\n".format(orpha, key))
# map orphanet rare diseases from orphanumber to doid
for line in orphaDisPhe_l:
if line.startswith('orphanet_code'):
continue
line_l = line.strip('\n').split('\t')
orpha_id = line_l[0]
#------------------------------------------------ orpha_name = line_l[1]
#----------------------------------------------------- hp_id = line_l[2]
#--------------------------------------------------- hp_name = line_l[3]
orpha_orpha_dct[orpha_id] = 1
if not orpha_id in orpha2do_dct:
mapping_cardinality = 0
disease_id = orpha_id
rd_sym_f.write('{}\t1:{}\t{}\t{}\t{}\t{}\n'.format(
line_l[0], mapping_cardinality, disease_id, line_l[1],
line_l[2], line_l[3]))
else:
doid_l = list(orpha2do_dct[orpha_id].keys())
mapping_cardinality = len(doid_l)
if len(doid_l) > 1:
print('orpha_id: {} has more than one doid'.format(orpha_id))
for disease_id in doid_l:
rd_sym_f.write('{}\t1:{}\t{}\t{}\t{}\t{}\n'.format(
line_l[0], mapping_cardinality, disease_id, line_l[1],
line_l[2], line_l[3]))
for orpha in orpha_orpha_dct:
if orpha in orpha2do_dct:
for doid in orpha2do_dct[orpha].keys():
orpha_mappings_f.write('{}\t{}\n'.format(orpha, doid))
# close files
rd_sym_f.close()
rd_mappings_f.close()
orpha_mappings_f.close()
do_orpha_set = set(do_orpha_dct.keys())
orpha_orpha_set = set(orpha_orpha_dct.keys())
print("")
print('From do, "xref" mappings DO to Orphanet:')
print('orphanet_orpha: {}; do_orpha: {} OrphaNumbers'.format(
len(orpha_orpha_set), len(do_orpha_set)))
commons = list(do_orpha_set & orpha_orpha_set)
differs = list(orpha_orpha_set - do_orpha_set)
print('mapped: {}'.format(len(commons)))
commons_h = commons[0:4]
print('mapped: {}'.format(commons_h))
print('not mapped: {}'.format(len(differs)))
differs_h = differs[0:4]
print('not mapped: {}'.format(differs_h))
def get_hpo_mesh_mappings():
'''
Function to infer hpo2mesh and mesh2hpo mappings from the UMLS metathesaurus
:return: hpo2mesh_dict
:return: mesh2hpo_dict
:return: hp2name_dict
:return: mesh2name_dict
'''
# VARIABLES
umls_hpo2umls_d = {}
umls_mesh2umls_d = {}
inferred_hpo2mesh_d = {}
inferred_mesh2hpo_d = {}
hp2name_d = {}
mesh2name_d = {}
# ALGORITHM
# Create hpo2umls dict from UMLS mappings
for line in open(
'/home/nuria/workspace/repurposing-hetio/rephetio-dhimmelstein/hetionet+hpo/data/umls-hpo-2016aa.tsv',
'r').readlines():
if line.startswith('CUI'):
continue
line_l = line.strip('\n').split('\t')
cui = line_l[0]
hp_code = line_l[2]
hp_term = line_l[3]
util.add_elem_with_dictionary(umls_hpo2umls_d, hp_code, cui)
hp2name_d[hp_code] = hp_term
#print('{}\t{}\t{}'.format(cui,hp_code,hp_term))
# Create mesh2umls dict from UMLS mappings
for line in open(
'/home/nuria/workspace/repurposing-hetio/rephetio-dhimmelstein/hetionet+hpo/data/umls-mesh-2016aa.tsv',
'r').readlines():
if line.startswith('CUI'):
continue
line_l = line.strip('\n').split('\t')
cui = line_l[0]
mesh_code = line_l[2]
mesh_term = line_l[3]
util.add_elem_with_dictionary(umls_mesh2umls_d, mesh_code, cui)
mesh2name_d[mesh_code] = mesh_term
#print('{}\t{}\t{}'.format(cui,mesh_code,mesh_term))
# Infer and create hpo2mesh dict
for hp in umls_hpo2umls_d:
hpCuis_s = set(umls_hpo2umls_d[hp])
for mesh in umls_mesh2umls_d:
meshCuis_s = set(umls_mesh2umls_d[mesh])
if len(hpCuis_s) > len(meshCuis_s) or len(hpCuis_s) == len(
meshCuis_s):
resta = hpCuis_s - meshCuis_s
else:
resta = meshCuis_s - hpCuis_s
if len(resta) == 0:
util.add_elem_with_dictionary(inferred_hpo2mesh_d, hp, mesh)
util.add_elem_with_dictionary(inferred_mesh2hpo_d, mesh, hp)
print('{} {} {} {} {} {} {}'.format(resta, hp, hp2name_d[hp],
hpCuis_s, mesh,
mesh2name_d[mesh],
meshCuis_s))
return inferred_hpo2mesh_d, inferred_mesh2hpo_d, hp2name_d, mesh2name_d