def mock_proteins_and_genes(count):
from database import db
from models import Gene, Protein
for i in range(count):
g = Gene(name='Gene_%s' % i, full_name='Full name of gene %s' % i)
p = Protein(refseq='NM_000%s' % i, gene=g)
g.preferred_isoform = p
db.session.add(g)
def test_autocomplete_all(self):
# MC3 GeneList is required as a target (a href for links) where users will be pointed
# after clicking of cancer autocomplete suggestion. Likewise with the ClinVar list.
db.session.add_all([
GeneList(name=name, mutation_source_name=detail_class.name)
for name, detail_class in [
('TCGA', MC3Mutation), ('ClinVar', InheritedMutation)
]
])
g = Gene(name='BR')
p = Protein(id=1, refseq='NM_007', gene=g, sequence='XXXXXV')
g.preferred_isoform = p # required for gene search to work - genes without preferred isoforms are ignored
mut = Mutation(protein=p, position=6, alt='E')
db.session.add_all([mut, p, g])
def autocomplete(query):
r = self.client.get('/search/autocomplete_all/?q=' + query)
self.visit_returned_urls(r)
return r
from database import bdb_refseq, bdb
bdb_refseq['BR V6E'] = [p.id] # required for mutation search
bdb.add_genomic_mut('1', 10000, 'T', 'C', mut)
# Gene and mutations
response = autocomplete('BR V6E')
entry = get_entry_and_check_type(response, 'aminoacid mutation')
assert entry
response = autocomplete('BR V6')
entry = get_entry_and_check_type(response, 'message')
assert 'Awaiting for <code>{alt}</code>' in entry['name']
response = autocomplete('BR V')
entry = get_entry_and_check_type(response, 'message')
assert 'Awaiting for <code>{pos}{alt}</code>' in entry['name']
response = autocomplete('B')
entry = get_entry_and_check_type(response, 'gene')
assert 'BR' == entry['name']
# genomic mutation
response = autocomplete('chr1 10000 T C')
entry = get_entry_and_check_type(response, 'nucleotide mutation')
assert entry and entry['input'] == 'CHR1 10000 T C'
# is the search falling back to the other strand?
response = autocomplete('chr1 10000 A G')
entry = get_entry_and_check_type(response, 'nucleotide mutation')
assert entry and entry['input'] == 'complement of CHR1 10000 A G'
prompt = 'Awaiting for mutation in <code>{chrom} {pos} {ref} {alt}</code> format'
for prompt_invoking_query in ['chr1', 'chr1 ', 'chr1 40', 'chr1 40 ', 'chr1 40 T']:
response = autocomplete(prompt_invoking_query)
entry = get_entry_and_check_type(response, 'message')
assert entry['name'] == prompt
# Pathways
pathways = [
Pathway(description='Activation of RAS in B cells', reactome=1169092),
Pathway(description='abortive mitotic cell cycle', gene_ontology=33277),
Pathway(description='amacrine cell differentiation', gene_ontology=35881),
Pathway(description='amniotic stem cell differentiation', gene_ontology=97086)
]
db.session.add_all(pathways)
# test partial matching and Reactome id pathways search
for ras_activation_query in ['Activation', 'REAC:1', 'REAC:1169092']:
response = autocomplete(ras_activation_query)
entry = get_entry_and_check_type(response, 'pathway')
assert entry['name'].startswith('Activation of RAS in B cells')
# test Gene Ontology search:
response = autocomplete('GO:33')
go_pathway = get_entry_and_check_type(response, 'pathway')
assert go_pathway['name'] == 'abortive mitotic cell cycle (GO:33277)'
# check if multiple pathways are returned
response = autocomplete('differentiation')
assert len(response.json['entries']) == 2
# check if both genes an pathways are returned simultaneously
# there should be: a pathway ('a>b<ortive...') and the >B<R gene
response = autocomplete('b')
entries = response.json['entries']
names = [entry['name'] for entry in entries]
assert all([name in names for name in ['BR', 'abortive mitotic cell cycle']])
# check if "search more pathways" is displayed
response = autocomplete('cell') # cell occurs in all four of added pathways;
# as a limit of pathways shown is 3, we should get a "show more" link
links = entries_with_type(response, 'see_more')
assert len(links) == 1
assert links[0]['name'] == 'Show all pathways matching <i>cell</i>'
# test case insensitive text search
response = autocomplete('AMNIOTIC STEM')
pathways = entries_with_type(response, 'pathway')
assert len(pathways) == 1
assert pathways[0]['name'] == 'amniotic stem cell differentiation'
# Disease
disease_names = [
'Cystic fibrosis', 'Polycystic kidney disease 2',
'Frontotemporal dementia', 'Cataract, nuclear total'
]
diseases = {name: Disease(name=name) for name in disease_names}
db.session.add_all(diseases.values())
response = autocomplete('cystic')
cystic_matching = entries_with_type(response, 'disease')
# both 'Cystic fibrosis' and PKD2 should match
assert len(cystic_matching) == 2
# is comma containing disease name properly linked?
response = autocomplete('Cataract')
cataract = get_entry_and_check_type(response, 'disease')
assert cataract['name'] == 'Cataract, nuclear total'
# Gene mutation in disease
# test suggestions
response = autocomplete('cystic ')
entry = entries_with_type(response, 'message')[0]
assert re.match('Do you wish to search for (.*?) mutations\?', entry['name'])
# currently there are no mutations associated with any disease
# so the auto-completion should not return any results
response = autocomplete('cystic in ')
assert not response.json['entries']
# let's add a mutation
m = Mutation(protein=p, position=1, alt='Y')
bdb_refseq['BR X1Y'] = ['NM_007']
# note: sig_code is required here
data = ClinicalData(disease=diseases['Cystic fibrosis'], sig_code=1)
disease_mutation = InheritedMutation(mutation=m, clin_data=[data])
db.session.add_all([m, data, disease_mutation])
# should return '.. in BR' suggestion now.
for query in ['cystic in', 'cystic in ']:
response = autocomplete(query)
result = get_entry_and_check_type(response, 'disease_in_protein')
assert result['gene'] == 'BR'
assert result['name'] == 'Cystic fibrosis'
# both gene search and refseq search should yield the same, non-empty results
results = []
for query in ['cystic in BR', 'cystic in NM_007', 'cystic in 007']:
response = autocomplete(query)
result = get_entry_and_check_type(response, 'disease_in_protein')
results.append(result)
assert all(r == result for r in results) and result
def test_mutated_sites(self):
g = Gene(name='Gene X')
p = Protein(refseq='NM_007', sequence='ABCDEFGHIJKLMNOPQRSTUVWXYZ', gene=g)
g.preferred_isoform = p
glycosylation = SiteType(name='glycosylation')
sites = {
# ClinVar muts and TCGA muts but different, with total count = 5 (3 + 2)
'A': Site(position=1, residue='A', protein=p),
# ClinVar muts intersection TCGA muts, total count = 4 (2 + 2)
'K': Site(position=11, residue='K', protein=p),
# Only TCGA muts, total count = 3 (1 + 2)
'U': Site(position=21, residue='U', protein=p, types={glycosylation})
}
def mut(pos):
return Mutation(position=pos, alt='X', protein=p)
intersecting_mut = mut(11)
mutations = [
# the first site (1 A)
InheritedMutation(
mutation=mut(1),
clin_data=[ClinicalData(), ClinicalData(), ClinicalData()]
),
MC3Mutation(mutation=mut(2), count=2),
# the second site (11 K)
InheritedMutation(
mutation=intersecting_mut,
clin_data=[ClinicalData(), ClinicalData()]
),
MC3Mutation(mutation=intersecting_mut, count=2),
# the third site (21 U)
MC3Mutation(mutation=mut(20), count=1),
MC3Mutation(mutation=mut(22), count=2),
]
db.session.add_all(mutations)
db.session.add_all([p, g])
db.session.add_all(sites.values())
db.session.commit()
sites_with_clinvar = most_mutated_sites([InheritedMutation]).all()
assert sites_with_clinvar == [(sites['A'], 3), (sites['K'], 2)]
sites_with_mc3 = most_mutated_sites([MC3Mutation]).all()
assert set(sites_with_mc3) == {(sites['A'], 2), (sites['K'], 2), (sites['U'], 3)}
both_sources = [MC3Mutation, InheritedMutation]
sites_with_muts_in_both_intersection = most_mutated_sites(both_sources, intersection=True).all()
assert sites_with_muts_in_both_intersection == [(sites['K'], 4)]
sites_with_muts_in_both = most_mutated_sites(both_sources, intersection=False).all()
assert sites_with_muts_in_both == [(sites['A'], 5), (sites['K'], 4)]
glyco_sites_with_mc3 = most_mutated_sites([MC3Mutation], site_type=glycosylation).all()
assert glyco_sites_with_mc3 == [(sites['U'], 3)]
