def update_table(facet_dropdown, facet_range, facet_min, facet_max, dropdown_id, min_id, facets, graph, label):
print("Facet inputs", facet_dropdown, facet_range, facet_min, facet_max)
facet_inputs = {}
for f, v in zip(dropdown_id, facet_dropdown):
facet_inputs[f['index']] = v
for f, s, n, m in zip(min_id, facet_range, facet_min, facet_max):
facet_inputs[f['index']] = [s, n, m]
if graph is not None and label is not None:
conn = connect()
G = conn.graph(graph)
q = G.query().V().hasLabel(label)
if len(facet_dropdown):
for k, f in facets.items():
if f['type'] in ["STRING", "BOOL"]:
fi = facet_inputs[f['index']]
if len(fi):
fset = []
# turn the dropdown selection numbers back into the original values
for j in fi:
fset.append(f['values'][j])
q = q.has(gripql.within(k, fset))
elif f["type"] == "NUMERIC":
if len(facet_inputs[f['index']][0]) > 0:
q = q.has(gripql.inside(k, facet_inputs[f['index']][1], facet_inputs[f['index']][2]))
data = results_data(q)
return data, query_string(q)
return [], "V()"
def test_has_within(O):
errors = []
setupGraph(O)
count = 0
for i in O.query().V().has(gripql.within("occupation", ["jedi", "sith"])):
count += 1
if i['gid'] not in ["vertex2", "vertex5", "vertex6"]:
errors.append("Wrong vertex returned %s" % (i))
if count != 3:
errors.append(
"Fail: O.query().V().has(gripql.within(\"occupation\", [\"jedi\", \"sith\"])) %s != %s" %
(count, 3))
count = 0
for i in O.query().V().has(gripql.within("occupation", 0)):
count += 1
if count != 0:
errors.append(
"Fail: O.query().V().has(gripql.within(\"occupation\", 0)) %s != %s" %
(count, 0))
return errors
def find_drugs_for_mutation_dataset(self, genes, dataset):
program = "Program:" + dataset.upper()
q = self.O.query().V(program).in_("InProgram").in_("InProject").in_(
"SampleFor").in_("AliquotFor").distinct("_gid")
all_aliquots = []
for row in q:
all_aliquots.append(row.gid)
# GENES = ["CDKN2A", "PTEN", "TP53", "SMAD4"]
gene_ids = {}
for g in genes:
for i in self.O.query().V().hasLabel("Gene").has(
gripql.eq("symbol", g)):
gene_ids[g] = i.gid
#Scan <dataset> cell lines based on mutation status
mut_samples = {}
norm_samples = {}
q = self.O.query().V(all_aliquots).as_("sample").in_(
"CallsetFor").outE("AlleleCall")
q = q.has(gripql.within("ensembl_gene",
list(gene_ids.values()))).as_("variant")
q = q.render({
"sample": "$sample._gid",
"gene": "$variant._data.ensembl_gene"
})
for res in q:
mut_samples[res.gene] = mut_samples.get(res.gene, set()) | set(
[res.sample])
# get dataset samples without mutation
for i in gene_ids.values():
norm_samples[i] = list(
set(all_aliquots).difference(mut_samples[i]))
print("%s Positive Set: %d" % (i, len(mut_samples[i])))
print("%s Negative Set: %d" % (i, len(norm_samples[i])))
# Get response values for the positive set (samples with mutation) and collect AUC value by drug
pos_response = {}
compound = {}
for g in gene_ids.values():
pos_response[g] = {}
q = self.O.query().V(list(mut_samples[g])).in_("ResponseIn").has(
gripql.eq("source",
dataset)).as_("a").out("ResponseTo").as_("b").select(
["a", "b"])
for row in q:
if hasattr(row['a']['data'],
'act_area'): # not all rows have 'amax' attribute
v = row['a']['data']['act_area']
else:
v = 0
id = row['b']['gid']
compound[id] = row['b']['data']['name']
if id not in pos_response[g]:
pos_response[g][id] = [v]
else:
pos_response[g][id].append(v)
#Get response values for the negative set (samples without mutation) and collect AUC value by drug
neg_response = {}
for g in gene_ids.values():
neg_response[g] = {}
q = self.O.query().V(list(norm_samples[g])).in_("ResponseIn").has(
gripql.eq("source",
dataset)).as_("a").out("ResponseTo").as_("b").select(
["a", "b"])
for row in q:
if hasattr(row['a']['data'],
'act_area'): # not all rows have 'amax' attribute
v = row['a']['data']['act_area']
else:
v = 0
id = row['b']['gid']
compound[id] = row['b']['data']['name']
if id not in neg_response[g]:
neg_response[g][id] = [v]
else:
neg_response[g][id].append(v)
#Collect t-test statistics
drugs = set(
itertools.chain.from_iterable(i.keys()
for i in pos_response.values()))
out = []
for drug in drugs:
for g in gene_ids.values():
if drug in pos_response[g] and drug in neg_response[g]:
row = {"drug": drug, "mutation": g}
mut_values = pos_response[g][drug]
norm_values = neg_response[g][drug]
if len(mut_values) > 5 and len(norm_values) > 5:
s = stats.ttest_ind(mut_values,
norm_values,
equal_var=False)
if s.pvalue <= 0.05 and s.statistic > 0: # means drug is significantly effective
out.append(compound[drug])
# print(out)
# get names of compounds
return out
def find_drugs_for_mutation_dataset(self, genes, dataset):
dataset = dataset.upper()
program = "Program:" + dataset
q = self.O.query().V(program).out("projects").out("cases").distinct(
"_gid")
all_cases = []
for row in q:
all_cases.append(row.gid)
gene_ids = {}
for i in self.O.query().V().hasLabel("Gene").has(
gripql.within("symbol", genes)):
gene_ids[i.data.symbol] = i.gid
mut_cases = {}
norm_cases = {}
q = self.O.query().V(all_cases).as_("ds")
if dataset != "CCLE":
q = q.out("same_as").has(gripql.eq("project_id", "Project:CCLE"))
q = q.out("samples").out("aliquots").out("somatic_callsets")
q = q.outE("alleles").has(
gripql.within("ensembl_gene", list(gene_ids.values())))
q = q.render({"case": "$ds._gid", "gene": "$._data.ensembl_gene"})
for res in q:
mut_cases[res.gene] = mut_cases.get(res.gene, set()) | set(
[res.case])
#get CCLE samples without mutation
for i in gene_ids.values():
norm_cases[i] = list(set(all_cases).difference(mut_cases[i]))
print("%s Positive Set: %d" % (i, len(mut_cases[i])))
print("%s Negative Set: %d" % (i, len(norm_cases[i])))
names = {}
area_metric = "act_area" if dataset == "CCLE" else "auc"
area_metric = "aac"
pos_response = {}
for g in gene_ids.values():
pos_response[g] = {}
q = self.O.query().V(list(
mut_cases[g])).as_("a").out("samples").out("aliquots")
q = q.out("drug_response").as_("a").out("compounds").as_("b")
q = q.select(["a", "b"])
for row in q:
if hasattr(row["a"]["data"], area_metric):
v = row["a"]["data"][area_metric]
else:
v = 0
id = row["b"]["gid"]
names[id] = get_drug_name_from_compound(row["b"]["data"])
if id not in pos_response[g]:
pos_response[g][id] = [v]
else:
pos_response[g][id].append(v)
neg_response = {}
for g in gene_ids.values():
neg_response[g] = {}
q = self.O.query().V(list(
norm_cases[g])).as_("a").out("samples").out("aliquots")
q = q.out("drug_response").as_("a").out("compounds").as_("b")
q = q.select(["a", "b"])
for row in q:
if hasattr(row["a"]["data"], area_metric):
v = row["a"]["data"][area_metric]
else:
v = 0
id = row["b"]["gid"]
names[id] = get_drug_name_from_compound(row["b"]["data"])
if id not in neg_response[g]:
neg_response[g][id] = [v]
else:
neg_response[g][id].append(v)
drugs = set(
itertools.chain.from_iterable(i.keys()
for i in pos_response.values()))
out = []
for drug in drugs:
for g in gene_ids.values():
if drug in pos_response[g] and drug in neg_response[g]:
row = {"drug": drug, "mutation": g}
mut_values = pos_response[g][drug]
norm_values = neg_response[g][drug]
if len(mut_values) > 5 and len(norm_values) > 5:
s = stats.ttest_ind(mut_values,
norm_values,
equal_var=False)
if s.pvalue <= 0.05: # means drug is significantly effective
n = names[drug]
out.append(n)
return out
def within(self, *args, **kwargs):
return gripql.within(*args, **kwargs)