def test_get_molecular_function_description(self):
desc = QEBI.get_molecular_function_description('GO:0004689')
self.assertIsNotNone(desc)
self.assertEqual(desc, get_from_test_file('GO:0004689'))
desc = QEBI.get_molecular_function_description('GO:00046890')
self.assertEqual(desc, 'None')
def test_get_disease_description(self):
desc = QEBI.get_disease_description('OMIM:613573')
self.assertIsNotNone(desc)
self.assertEqual(desc, get_from_test_file('OMIM:613573'))
desc = QEBI.get_disease_description('OMIM:6135730')
self.assertEqual(desc, 'None')
def test_get_cellular_component_description(self):
desc = QEBI.get_cellular_component_description('GO:0005573')
self.assertIsNotNone(desc)
self.assertEqual(desc, get_from_test_file('GO:0005573'))
desc = QEBI.get_cellular_component_description('GO:00055730')
self.assertEqual(desc, 'None')
def test_get_bio_process_description(self):
desc = QEBI.get_bio_process_description('HP:0011105')
self.assertIsNotNone(desc)
self.assertEqual(desc, get_from_test_file('HP:0011105'))
desc = QEBI.get_bio_process_description('HP:00111050')
self.assertEqual(desc, 'None')
def test_get_phenotype_description(self):
desc = QEBI.get_phenotype_description('GO:0042535')
self.assertIsNotNone(desc)
self.assertEqual(desc, get_from_test_file('GO:0042535'))
desc = QEBI.get_phenotype_description('GO:00425350')
self.assertEqual(desc, 'None')
def test_get_anatomy_description(self):
desc = QEBI.get_anatomy_description('UBERON:0004476')
self.assertIsNotNone(desc)
self.assertEqual(desc, get_from_test_file('UBERON:0004476'))
desc = QEBI.get_anatomy_description('UBERON:00044760')
self.assertEqual(desc, 'None')
desc = QEBI.get_anatomy_description('CL:0000038')
self.assertIsNotNone(desc)
self.assertEqual(desc, get_from_test_file('CL:0000038'))
desc = QEBI.get_anatomy_description('CL:00000380')
self.assertEqual(desc, 'None')
def update_molecular_function_nodes_desc(self):
conn = Neo4jConnection(self.neo4j_url, self.neo4j_user,
self.neo4j_password)
nodes = conn.get_molecular_function_nodes()
print("the number of molecular_function nodes: %d" % len(nodes))
from time import time
t = time()
nodes_array = []
for i, node_id in enumerate(nodes):
# print("no %d" % i)
node = dict()
node['node_id'] = node_id
node['desc'] = QueryEBIOLS.get_molecular_function_description(
node_id)
nodes_array.append(node)
print("molecular_function pulling time: %f" % (time() - t))
nodes_nums = len(nodes_array)
chunk_size = 10000
group_nums = nodes_nums // chunk_size + 1
for i in range(group_nums):
start = i * chunk_size
end = (i + 1) * chunk_size if (
i + 1) * chunk_size < nodes_nums else nodes_nums
conn.update_molecular_function_nodes_desc(nodes_array[start:end])
print("molecular_function total time: %f" % (time() - t))
conn.close()
def update_disease_nodes_desc(self):
conn = Neo4jConnection(self.neo4j_url, self.neo4j_user,
self.neo4j_password)
nodes = conn.get_disease_nodes()
print("the number of disease nodes: %d" % len(nodes))
from time import time
t = time()
nodes_array = []
qo = QueryOMIM()
for i, node_id in enumerate(nodes):
node = dict()
node['node_id'] = node_id
if node_id[:4] == "OMIM":
node['desc'] = qo.disease_mim_to_description(node_id)
elif node_id[:4] == "DOID":
node['desc'] = QueryEBIOLS.get_disease_description(node_id)
nodes_array.append(node)
print("disease api pulling time: %f" % (time() - t))
nodes_nums = len(nodes_array)
chunk_size = 10000
group_nums = nodes_nums // chunk_size + 1
for i in range(group_nums):
start = i * chunk_size
end = (i + 1) * chunk_size if (
i + 1) * chunk_size < nodes_nums else nodes_nums
conn.update_disease_nodes_desc(nodes_array[start:end])
print("disease total time: %f" % (time() - t))
conn.close()
def test_update_phenotype_nodes_desc(self):
conn = Neo4jConnection(self.rtxConfig.neo4j_bolt,
self.rtxConfig.neo4j_username,
self.rtxConfig.neo4j_password)
nodes = conn.get_phenotype_nodes()
# generate random number array
random_indexes = random_int_list(0, len(nodes) - 1, 100)
for i in random_indexes:
# retrieve data from API
node_id = nodes[i]
desc = QueryEBIOLS.get_phenotype_description(node_id)
# retrieve data from Neo4j
node = conn.get_phenotype_node(node_id)
self.assertIsNotNone(node)
self.assertIsNotNone(node['n']['id'])
self.assertIsNotNone(node['n']['description'])
self.assertEqual(node_id, node['n']['id'])
if node['n']['description'] != "None":
self.assertEqual(desc, node['n']['description'])
conn.close()
def test_update_phenotype_nodes_desc(self):
f = open('config.json', 'r')
config_data = f.read()
f.close()
config = json.loads(config_data)
conn = Neo4jConnection(config['url'], config['username'], config['password'])
nodes = conn.get_phenotype_nodes()
# generate random number array
random_indexes = random_int_list(0, len(nodes)-1, 100)
for i in random_indexes:
# retrieve data from API
node_id = nodes[i]
desc = QueryEBIOLS.get_phenotype_description(node_id)
# retrieve data from Neo4j
node = conn.get_phenotype_node(node_id)
self.assertIsNotNone(node)
self.assertIsNotNone(node['n']['id'])
self.assertIsNotNone(node['n']['description'])
self.assertEqual(node_id, node['n']['id'])
if node['n']['description'] != "None":
self.assertEqual(desc, node['n']['description'])
conn.close()
def get_mesh_term_for_all(curie_id, description):
"""
Takes a curie ID, detects the ontology from the curie id, and then finds the mesh term
Params:
curie_id - A string containing the curie id of the node. Formatted <source abbreviation>:<number> e.g. DOID:8398
description - A string containing the English name for the node
current functionality (+ means has it, - means does not have it)
"Reactome" +
"GO" - found gene conversion but no biological process conversion
"UniProt" +
"HP" - +
"UBERON" +
"CL" - not supposed to be here?
"NCBIGene" +
"DOID" +
"OMIM" +
"ChEMBL" +
"""
if type(description) != str:
description = str(description)
curie_list = curie_id.split(':')
names = None
if QueryNCBIeUtils.is_mesh_term(description):
return [description + '[MeSH Terms]']
names = NormGoogleDistance.get_mesh_from_oxo(curie_id)
if names is None:
if curie_list[0].lower().startswith("react"):
res = QueryNCBIeUtils.get_reactome_names(curie_list[1])
if res is not None:
names = res.split('|')
elif curie_list[0] == "GO":
pass
elif curie_list[0].startswith("UniProt"):
res = QueryNCBIeUtils.get_uniprot_names(curie_list[1])
if res is not None:
names = res.split('|')
elif curie_list[0] == "HP":
names = QueryNCBIeUtils.get_mesh_terms_for_hp_id(curie_id)
elif curie_list[0] == "UBERON":
if curie_id.endswith('PHENOTYPE'):
curie_id = curie_id[:-9]
mesh_id = QueryEBIOLS.get_mesh_id_for_uberon_id(curie_id)
names = []
for entry in mesh_id:
if len(entry.split('.')) > 1:
uids=QueryNCBIeUtils.get_mesh_uids_for_mesh_tree(entry.split(':')[1])
for uid in uids:
try:
uid_num = int(uid.split(':')[1][1:]) + 68000000
names += QueryNCBIeUtils.get_mesh_terms_for_mesh_uid(uid_num)
except IndexError:
uid_num = int(uid)
names += QueryNCBIeUtils.get_mesh_terms_for_mesh_uid(uid_num)
else:
try:
uid = entry.split(':')[1]
uid_num = int(uid[1:]) + 68000000
names += QueryNCBIeUtils.get_mesh_terms_for_mesh_uid(uid_num)
except IndexError:
uid_num = int(entry)
names += QueryNCBIeUtils.get_mesh_terms_for_mesh_uid(uid_num)
if len(names) == 0:
names = None
else:
names[0] = names[0] + '[MeSH Terms]'
elif curie_list[0] == "NCBIGene":
gene_id = curie_id.split(':')[1]
names = QueryNCBIeUtils.get_pubmed_from_ncbi_gene(gene_id)
elif curie_list[0] == "DOID":
mesh_id = QueryDisont.query_disont_to_mesh_id(curie_id)
names = []
for uid in mesh_id:
uid_num = int(uid[1:]) + 68000000
name = QueryNCBIeUtils.get_mesh_terms_for_mesh_uid(uid_num)
if name is not None:
names += name
if len(names) == 0:
names = None
else:
names[0] = names[0] + '[MeSH Terms]'
elif curie_list[0] == "OMIM":
names = QueryNCBIeUtils.get_mesh_terms_for_omim_id(curie_list[1])
elif curie_list[0] == "ChEMBL":
chembl_id = curie_id.replace(':', '').upper()
mesh_id = QueryMyChem.get_mesh_id(chembl_id)
if mesh_id is not None:
mesh_id = int(mesh_id[1:]) + 68000000
names = QueryNCBIeUtils.get_mesh_terms_for_mesh_uid(mesh_id)
if names is not None:
if type(names) == list:
for name in names:
if name.endswith('[MeSH Terms]'):
return [name]
return names
return [description.replace(';', '|')]
def setUpClass(cls):
cls.ebiols = QueryEBIOLS()
def query_neighbor_genes_for_gene_set_in_a_given_anatomy(self,
anatomy_curie_id_str,
protein_set_curie_id_str):
assert type(protein_set_curie_id_str) == tuple
assert len(protein_set_curie_id_str) > 0
assert type(anatomy_curie_id_str) == str
# convert UBERON anatomy curie ID str to a brenda anatomy ID
assert anatomy_curie_id_str.startswith("UBERON:")
bto_id_set = QueryEBIOLS.get_bto_id_for_uberon_id(anatomy_curie_id_str)
ret_dict = dict()
if len(bto_id_set) == 0:
return ret_dict
assert len(bto_id_set) == 1
bto_term = QueryEBIOLS.get_bto_term_for_bto_id(next(iter(bto_id_set))).replace(" ", "_")
entrez_gene_ids = set()
entrez_gene_ids_int = set()
# convert uniprot IDs to Entrez gene IDs
for protein_curie_id_str in protein_set_curie_id_str:
assert protein_curie_id_str.startswith("UniProtKB:")
uniprot_acc = protein_curie_id_str.split(":")[1]
entrez_gene_id_set = self.mg.convert_uniprot_id_to_entrez_gene_ID(uniprot_acc)
for entrez_gene_id in entrez_gene_id_set:
entrez_gene_ids_int.add(entrez_gene_id)
entrez_gene_ids.add(str(entrez_gene_id))
entrez_gene_ids_str = ",".join(entrez_gene_ids)
data = {"ids": entrez_gene_ids_str,
"tissue": bto_term,
"limit": self.limit}
results = self._wrapper(self.ENDPOINT, data)
ret_dict = dict()
gene_dict = dict()
for index, row in results.iterrows():
gene1 = row["Gene1"]
gene2 = row["Gene2"]
avg_corr = row["aveCorr"]
assert type(gene1) == int
assert type(gene2) == int
assert type(avg_corr) == float
if gene1 in entrez_gene_ids_int:
if gene2 in entrez_gene_ids_int:
# do nothing since this is not a new gene
new_gene_id = None
else:
# gene2 is the new gene
new_gene_id = gene2
else:
if gene2 in entrez_gene_ids_int:
new_gene_id = gene1
else:
print("neither gene was in the set of query genes, this should not happen", file=sys.stderr)
assert False
if new_gene_id is not None:
gene_dict[new_gene_id] = avg_corr
for gene_id, avg_corr in gene_dict.items():
uniprot_id_set = self.mg.convert_entrez_gene_id_to_uniprot_id(gene_id)
if len(uniprot_id_set) > 0:
for uniprot_id in uniprot_id_set:
ret_dict["UniProtKB:" + uniprot_id] = avg_corr
query_res = get_nodes_that_match_in_list(ret_dict.keys(), 'protein')
res_list = str(query_res[0])
res_list = ast.literal_eval(res_list[res_list.find('['):-1])
for uniprot_id in list(ret_dict):
if uniprot_id not in res_list:
ret_dict.pop(uniprot_id)
return ret_dict