def test_divide_muts_by_sites(self):
from views.network import divide_muts_by_sites
# check if null case works
divide_muts_by_sites([], [])
# one site
s_1 = Site(position=1)
muts_by_sites = divide_muts_by_sites([], [s_1])
assert muts_by_sites[s_1] == []
# full test
s_2 = Site(position=10)
s_3 = Site(position=20)
muts_by_pos = {
pos: [Mutation(position=pos)]
for pos in (1, 2, 8, 14, 16, 30)
}
muts_by_pos[16].append(Mutation(position=16))
def get_muts_from_pos(*positions):
lists = [muts_by_pos[p] for p in positions]
return [mut for mut_list in lists for mut in mut_list]
muts_by_sites = divide_muts_by_sites([
mut for muts_on_pos_x in muts_by_pos.values()
for mut in muts_on_pos_x
], [s_1, s_2, s_3])
assert muts_by_sites[s_1] == get_muts_from_pos(1, 2, 8)
assert muts_by_sites[s_2] == get_muts_from_pos(8, 14, 16)
assert muts_by_sites[s_3] == get_muts_from_pos(14, 16)
def get_filter_by_sources(sources):
filters = and_(((Mutation.get_relationship(source).any()
if are_details_managed(source) else
Mutation.get_relationship(source).has())
for source in sources))
return filters
def create_network():
p = create_test_protein()
cancer = Cancer(name='Ovarian', code='OV')
known_interactor_of_x = create_test_kinase('Kinase Y', 'NM_0009')
kinase_mutation = Mutation(position=1,
alt='T',
meta_MC3=[MC3Mutation(cancer=cancer)])
known_interactor_of_x.protein.mutations = [kinase_mutation]
drug = Drug(
name='Drug targeting ' + known_interactor_of_x.name,
drug_bank_id='DB01',
target_genes=[known_interactor_of_x.protein.gene],
# by default only approved drugs are shown
groups={DrugGroup(name='approved')})
group = KinaseGroup(name='Group of kinases', )
s = Site(position=1,
type='phosphorylation',
residue='T',
kinases=[known_interactor_of_x],
kinase_groups=[group])
s2 = Site(position=2,
type='phosphorylation',
residue='R',
kinase_groups=[group])
p.sites = [s, s2]
predicted_interactor = create_test_kinase('Kinase Z', 'NM_0002')
protein_mutation = Mutation(position=2,
alt='T',
meta_MC3=[MC3Mutation(cancer=cancer)],
meta_MIMP=[
MIMPMutation(
pwm=known_interactor_of_x.name,
effect='loss',
site=s,
probability=0.1,
position_in_motif=1),
MIMPMutation(pwm=predicted_interactor.name,
effect='gain',
site=s,
probability=0.1,
position_in_motif=1)
])
p.mutations = [protein_mutation]
db.session.add_all([p, drug, predicted_interactor])
db.session.commit()
# a new cancer was added, reload is necessary (this should not happen during normal app usage)
from website.views.filters import cached_queries
cached_queries.reload()
def test_search_mutations(self):
s = Site(position=13, types={SiteType(name='methylation')})
p = Protein(refseq='NM_007', id=7, sites=[s], sequence='XXXXXXXXXXXXV')
m_in_site = Mutation(protein=p, position=13, alt='V')
m_out_site = Mutation(protein=p, position=50, alt='K')
db.session.add(p)
# points to the same location as first record in VCF_FILE_CONTENT
test_query = 'chr20 14370 G A'
from database import bdb
# map the first genomic mutation from VCF_FILE_CONTENT
# to some (mocked) protein mutation
bdb.add_genomic_mut('20', 14370, 'G', 'A', m_in_site, is_ptm=True)
#
# basic test - is appropriate mutation in results?
#
response = self.search_mutations(mutations=test_query)
assert response.status_code == 200
# this mutation is exactly at a PTM site and should be included in results
assert '<td>{0}</td>'.format(m_in_site.alt).encode() in response.data
# this mutation lies outside of a PTM site - be default should be filtered out
assert '<td>{0}</td>'.format(m_out_site.alt).encode() not in response.data
#
# count test - is mutation for this query annotated as shown twice?
#
response = self.search_mutations(
mutations='{0}\n{0}'.format(test_query)
)
assert response.status_code == 200
assert b'<td>2</td>' in response.data
#
# VCF file test
#
response = self.client.post(
'/search/mutations',
content_type='multipart/form-data',
data={
'vcf-file': (BytesIO(VCF_FILE_CONTENT), 'exemplar_vcf.vcf')
}
)
assert response.status_code == 200
assert b'NM_007' in response.data
def test_prepare_dataset(self):
from views.mutation import prepare_datasets
p = Protein(refseq='NM_000123',
sequence='TRAN',
gene=Gene(name='TP53'))
mutation = Mutation(protein=p, position=2, alt='K')
details = MC3Mutation(mutation=mutation, count=2)
db.session.add(mutation)
datasets, user_datasets = prepare_datasets(mutation)
expected_datasets = [{
'filter': 'Mutation.sources:in:' + source.name,
'name': source.display_name,
'mutation_present': False
} if source is not MC3Mutation else {
'filter': 'Mutation.sources:in:' + MC3Mutation.name,
'name': MC3Mutation.display_name,
'mutation_present': [details]
} for source in source_manager.confirmed]
assert datasets == expected_datasets
assert not user_datasets
def test_sites(self):
mutations = [Mutation(position=x) for x in (0, 5, 12, 57)]
protein = Protein(refseq='NM_00002',
mutations=mutations,
sites=[Site(position=x) for x in (10, 14, 15, 57)])
db.session.add(protein)
db.session.commit()
# ==test_find_closest_sites==
# for mutation at position 0 there is no closest site;
# for mutation at position 5 there should be 1 closest site
expected_closest_sites = dict(zip(mutations, [0, 1, 2, 1]))
for mutation, expected_sites_cnt in expected_closest_sites.items():
sites_found = mutation.find_closest_sites()
assert len(sites_found) == expected_sites_cnt
# ==test_get_affected_ptm_sites==
expected_affected_sites = dict(zip(mutations, [0, 1, 3, 1]))
for mutation, expected_sites_cnt in expected_affected_sites.items():
sites_found = mutation.get_affected_ptm_sites()
assert len(sites_found) == expected_sites_cnt
def count_by_sources(sources: List[MutationSource],
site_type: SiteType,
primary_isoforms=True,
by_genes=True,
genes=None,
muts_conjunction=or_,
**kwargs):
base_query = Mutation.query.filter(
muts_conjunction(*[Mutation.in_sources(source) for source in sources]))
if primary_isoforms:
base_query = base_query.join(Protein).filter(
Protein.is_preferred_isoform)
sites = Site.query.filter(Site.types.contains(site_type))
counter = MotifsCounter(site_type)
if not by_genes:
return counter.count_muts_and_sites(base_query, sites, **kwargs)
counts_by_genes = {}
if not genes:
genes = Gene.query.all()
for gene in tqdm(genes):
query = base_query.filter(Mutation.protein == gene.preferred_isoform)
gene_sites = sites.filter(Site.protein == gene.preferred_isoform)
counts_by_genes[gene.name] = counter.count_muts_and_sites(
query, gene_sites, show_progress=False, **kwargs)
return counts_by_genes
def create_test_models():
protein = Protein(refseq='NM_0001',
gene=Gene(name='SOMEGENE'),
sequence='ABCD')
mutation = Mutation(protein=protein, position=1, alt='E')
protein.gene.preferred_isoform = protein
MC3Mutation(mutation=mutation,
cancer=Cancer(code='CAN'),
samples='Sample A,Sample B',
count=2)
InheritedMutation(mutation=mutation,
clin_data=[
ClinicalData(disease=Disease(name='Some disease'),
sig_code=5),
ClinicalData(disease=Disease(name='Other disease'),
sig_code=2)
])
protein_kinase = Protein(refseq='NM_0002',
gene=Gene(name='OTHERGENE'),
sequence='ABCD')
kinase = Kinase(name='Kinase name', protein=protein_kinase)
site = Site(protein=protein,
position=1,
residue='A',
kinases={kinase},
pmid={1, 2},
types={SiteType(name='glycosylation')})
protein.sites = [site]
return locals()
def mutation_sources():
sources = {}
for name, source in Mutation.sources_dict.items():
if name == 'user':
continue
sources[name] = Mutation.get_source_model(name)
return sources
def source_specific_mutated_sites():
muts_in_ptm_sites = {}
mimp_muts = {}
mutated_sites = defaultdict(dict)
site_type_queries = [models.SiteType(name='')] # empty will match all sites
site_type_queries.extend(models.SiteType.query)
for name, model in mutation_sources().items():
count = (
Mutation.query
.filter_by(is_confirmed=True, is_ptm_distal=True)
.filter(Mutation.in_sources(model))
.count()
)
muts_in_ptm_sites[name] = count
mimp_muts[name] = (
Mutation.query
.filter(
and_(
Mutation.in_sources(models.MIMPMutation, model),
Mutation.is_confirmed,
)
).count()
)
for site_type in tqdm(site_type_queries):
mutated_sites[name][site_type] = count_mutated_sites([site_type], model)
all_mutated_sites = {}
for site_type in tqdm(site_type_queries):
all_mutated_sites[site_type] = count_mutated_sites([site_type])
mutated_sites['merged'] = all_mutated_sites
return {
'Mutations - in PTM sites': muts_in_ptm_sites,
'Mutations - with network-rewiring effect': mimp_muts,
'PTM sites affected by mutations': mutated_sites
}
def with_significant_genes(self, significant_gene_list_name):
query = request.args.get('query', '')
gene_list = GeneList.query.filter_by(
name=significant_gene_list_name).first_or_404()
dataset = Mutation.get_source_model(gene_list.mutation_source_name)
return template('pathway/significant.html',
gene_list=gene_list,
dataset=dataset,
endpoint='significant_data',
endpoint_kwargs={'gene_list_id': gene_list.id},
query=query)
def get_genes_with_mutations_from_sources(sources,
only_genes_with_ptm_sites=False):
query = (db.session.query(Gene).join(
Protein, Gene.preferred_isoform_id == Protein.id).join(Mutation))
query = query.filter(Mutation.in_sources(*sources))
genes = set(query.distinct())
if only_genes_with_ptm_sites:
return {gene for gene in genes if gene.preferred_isoform.sites}
return genes
def mutated_ptm_sites_in_proximity(mutation_source,
type_1: str,
type_2: str,
mutation_filter=True,
distance: int = 7,
only_preferred=True) -> int:
sites, (site_1, site_2) = ptm_sites_in_proximity(type_1, type_2, distance,
only_preferred)
for site in (site_1, site_2):
sites = sites.filter(
site.affected_by_mutations.any(
and_(Mutation.in_sources(mutation_source), mutation_filter)))
return sites
def most_mutated_sites(sources: List[MutationSource],
site_type: SiteType = None,
limit=25,
intersection=True,
exclusive=None,
mutation_filter=None):
"""Sources must have the same value_type (counts/frequencies)"""
assert not (intersection and exclusive)
counts = prepare_for_summing(sources)
query = (db.session.query(
Site, *[count.label(f'count_{i}')
for i, count in enumerate(counts)]).select_from(Mutation))
if intersection:
for source in sources:
query = query.join(source)
else:
for source in sources:
query = query.outerjoin(source)
if exclusive:
query = query.filter(~Mutation.in_sources(*exclusive))
if mutation_filter is not None:
query = query.filter(mutation_filter)
query = (query.join(Mutation.affected_sites).filter(
Site.protein.has(Protein.is_preferred_isoform)))
if site_type:
query = query.filter(SiteType.fuzzy_filter(site_type, join=True))
query = (query.group_by(Site).having(and_(*counts)))
query = query.subquery()
total_muts_count = reduce(
operator.add,
[getattr(query.c, f'count_{i}') for i in range(len(counts))])
total_muts_count = total_muts_count.label('mutations_count')
query = (db.session.query(
aliased(Site, query),
total_muts_count,
).order_by(desc(total_muts_count)))
return query.limit(limit)
def test_mutate(self):
p = Protein(sequence='ABCDE')
s = Site(protein=p, position=3, residue='C')
cases = {
1: 'XBCDE',
3: 'ABXDE',
}
for position, expected_seq in cases.items():
m = Mutation(protein=p, position=position, alt='X')
assert mutate_sequence(s, m, offset=2) == expected_seq
def test_show(self):
p = Protein(refseq='NM_000123',
sequence='TRAN',
gene=Gene(name='TP53'))
mutation = Mutation(protein=p, position=2, alt='K')
db.session.add(mutation)
response = self.client.get('/mutation/show/NM_000123/2/K')
assert response.status_code == 200
assert b'TP53' in response.data
assert b'NM_000123' in response.data
def test_impact_on_ptm(self):
mutations = [Mutation(position=61)]
protein = Protein(refseq='NM_00001', mutations=mutations)
db.session.add(protein)
protein.sites = [
Site(position=61),
Site(position=54),
Site(position=51)
]
mutation = mutations[0]
assert mutation.impact_on_ptm() == 'direct'
def get_confirmed_mutations(sources,
only_preferred=True,
genes=None,
confirmed_by_definition=False,
base_query=None):
"""
Utility to generate a query for retrieving confirmed mutations having specific mutation details.
Args:
sources: list of mutation details (sources) to be used to filter
the mutations (including sources with non-confirmed mutations)
only_preferred: include only mutations from preferred isoforms
genes: limit to genes from provided list
confirmed_by_definition: do not apply the expensive is_confirmed=True
filter as all sources include only confirmed mutations
base_query: the initial mutation query (allows to adjust selected columns)
Returns:
Query object yielding mutations.
"""
if not base_query:
base_query = Mutation.query
mutations = base_query
def only_from_primary_isoforms(mutations_query):
mutations_query = join_unique(mutations_query, Protein)
return mutations_query.filter(Protein.is_preferred_isoform)
if not confirmed_by_definition:
mutations = mutations.filter_by(is_confirmed=True)
# TODO: remove?
mutations = only_from_primary_isoforms(mutations)
if genes:
mutations = mutations.filter(
Protein.id.in_([g.preferred_isoform_id for g in genes]))
selected_mutations = mutations.filter(Mutation.in_sources(*sources))
if only_preferred:
selected_mutations = only_from_primary_isoforms(selected_mutations)
return selected_mutations
def count_mutated_sites(
site_types: Iterable[models.SiteType]=tuple(), model=None,
only_primary=False, disordered=None, custom_filter=None
):
filters = [
Mutation.protein_id == Protein.id,
Site.protein_id == Protein.id,
Mutation.precomputed_is_ptm
]
for site_type in site_types:
filters.append(models.SiteType.fuzzy_filter(site_type))
if custom_filter is not None:
filters.append(custom_filter)
if disordered is not None:
filters.append(Site.in_disordered_region == disordered)
query = (
db.session.query(
func.count(distinct(case(
[
(
(
Mutation.position.between(
Site.position - 7,
Site.position + 7
)
),
Site.id
)
],
else_=literal_column('NULL')
)))
)
.filter(and_(*filters))
.join(Mutation, Site.protein_id == Mutation.protein_id)
)
if model:
query = query.filter(Mutation.in_sources(model))
else:
query = query.filter(Mutation.is_confirmed == True)
if only_primary:
query = query.join(Protein).filter(Protein.is_preferred_isoform)
return query.scalar()
def mutation_by_source(combination,
site_type=None,
only_within_ptm_sites=False,
only_primary=False):
query = (Mutation.query.filter(Mutation.in_sources(*combination)))
if only_within_ptm_sites:
# query = query.filter(Mutation.is_ptm_distal == True)
query = query.filter(Mutation.precomputed_is_ptm)
if site_type:
query = query.filter(
Mutation.affected_sites.any(Site.types.contains(site_type)))
if only_primary:
query = query.join(Protein).filter(Protein.is_preferred_isoform)
return query.count()
def count_mutations_from_genes(genes,
sources,
only_preferred_isoforms=False,
strict=True):
"""Counts mutations and PTM mutations from isoforms from given set of genes.
Args:
genes: a list of Gene
only_preferred_isoforms: should only one isoform per gene
(the preferred/primary one) be used when filtering mutations?
sources: a list of MutationDetails - only confirmed mutations from
sources identified by given MutationDetail classes will be counted
"""
all_mutations_count = 0
ptm_mutations_count = 0
if strict:
base_query = (db.session.query(
Mutation.position, Mutation.alt,
Protein.id).select_from(Mutation).join(Protein))
else:
base_query = Mutation.query
for gene in tqdm(genes):
if only_preferred_isoforms:
proteins = [gene.preferred_isoform]
else:
proteins = gene.isoforms
mutations_filters = and_(
Mutation.protein_id.in_([p.id for p in proteins]),
Mutation.is_confirmed == True, Mutation.in_sources(*sources))
all_mutations_count += (
base_query.filter(mutations_filters).distinct().count())
ptm_mutations_count += (base_query.filter(
and_(Mutation.precomputed_is_ptm,
mutations_filters)).distinct().count())
print(all_mutations_count, ptm_mutations_count,
ptm_mutations_count / all_mutations_count)
return all_mutations_count, ptm_mutations_count
def test_browse_list(self):
from miscellaneous import make_named_temp_file
from test_imports.test_gene_list import raw_gene_list
from imports.protein_data import active_driver_gene_lists as load_active_driver_gene_lists
filename = make_named_temp_file(raw_gene_list)
# create gene list and genes
with self.app.app_context():
from imports.protein_data import ListData
gene_lists = load_active_driver_gene_lists(lists=(
ListData(name='TCGA', path=filename, mutations_source=TCGAMutation),
))
db.session.add_all(gene_lists)
# create preferred isoforms for genes
for i, gene in enumerate(Gene.query.all()):
# at least one mutation is required for gene on a gene list to be displayed
mut = Mutation()
MC3Mutation(mutation=mut)
p = Protein(refseq='NM_000%s' % i, mutations=[mut])
gene.isoforms = [p]
gene.preferred_isoform = p
# check the static template
response = self.client.get('/gene/list/TCGA')
assert response.status_code == 200
assert b'TCGA' in response.data
# check the dynamic data
response = self.client.get('/gene/list_data/TCGA?order=asc')
assert response.status_code == 200
gene_list = GeneList.query.filter_by(name='TCGA').one()
# all results retrieved
assert response.json['total'] == len(gene_list.entries)
# properly sorted by fdr
fdrs = [row['fdr'] for row in response.json['rows']]
assert fdrs == sorted(fdrs)
def create_test_models():
protein = Protein(refseq='NM_0001',
gene=Gene(name='SOMEGENE'),
sequence='ABCD')
mutation = Mutation(protein=protein, position=1, alt='E')
MC3Mutation(mutation=mutation,
cancer=Cancer(code='CAN'),
samples='Some sample')
InheritedMutation(
mutation=mutation,
clin_data=[ClinicalData(disease=Disease(name='Some disease'))])
protein_kinase = Protein(refseq='NM_0002',
gene=Gene(name='OTHERGENE'),
sequence='ABCD')
kinase = Kinase(name='Kinase name', protein=protein_kinase)
site = Site(protein=protein, position=1, residue='A', kinases=[kinase])
protein.sites = [site]
return locals()
def __init__(self,
protein,
filter_manager,
include_kinases_from_groups=False):
super().__init__(protein, filter_manager, include_kinases_from_groups)
sites, kinases, kinase_groups = self.get_sites_and_kinases(
only_sites_with_kinases=False)
tracks = prepare_tracks(protein, self.protein_mutations)
source = filter_manager.get_value('Mutation.sources')
source_model = Mutation.get_source_model(source)
value_type = source_model.value_type
parsed_mutations = self.represent_needles()
self.json_data = {
'value_type': value_type,
'log_scale': (value_type == 'frequency'),
'mutations': parsed_mutations,
'sites': prepare_sites(sites),
'tracks': tracks
}
def test_impact_on_specific_ptm(self):
# case 0: there are no sites in the protein
mutations = {
Mutation(position=10): 'none', # too far away
Mutation(position=9): 'none',
Mutation(position=8): 'distal',
Mutation(position=4): 'distal',
Mutation(position=3): 'proximal',
Mutation(position=2): 'proximal',
Mutation(position=1): 'direct'
}
protein = Protein(refseq='NM_00001', mutations=mutations.keys())
db.session.add(protein)
for mutation in mutations.keys():
assert mutation.impact_on_ptm() == 'none'
# case 1: there are some sites in the protein
protein.sites = [Site(position=1), Site(position=50)]
site = protein.sites[0]
for mutation, impact in mutations.items():
print(mutation)
assert mutation.impact_on_ptm() == impact
assert mutation.impact_on_specific_ptm(site) == impact
# case 2: there are some sites but all will be excluded by a site filter
def site_filter(sites):
return []
for mutation in mutations.keys():
assert mutation.impact_on_ptm(site_filter=site_filter) == 'none'
def count_by_sources(sources: List[MutationSource]):
return Mutation.query.filter(Mutation.in_sources(*sources)).count()
def confirmed_with_mimp(self):
return Mutation.query.filter(
and_(
Mutation.in_sources(models.MIMPMutation),
Mutation.is_confirmed,
)).count()
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 source_specific_nucleotide_mappings():
from database import bdb
from genomic_mappings import decode_csv
from models import Mutation
from tqdm import tqdm
from gc import collect
mutations = defaultdict(str)
def count_mutations(mutations_query):
for mutation in tqdm(mutations_query, total=mutations_query.count()):
mutations[str(mutation[0]) + mutation[1] + str(mutation[2])] += i
sources_map = {str(i): model for i, model in enumerate(mutation_sources().values())}
print('Loading mutations from sources:')
for i, model in tqdm(sources_map.items(), total=len(sources_map)):
query = (
db.session.query(Mutation.protein_id, Mutation.alt, Mutation.position)
.filter(Mutation.in_sources(model))
# no need for '.filter(Mutation.is_confirmed==True)'
# (if it is in source of interest, it is confirmed - we do not count MIMPs here)
.yield_per(5000)
)
count_mutations(query)
# add merged
i = str(len(sources_map))
sources_map[i] = 'merged'
print('Loading merged mutations:')
query = (
db.session.query(Mutation.protein_id, Mutation.alt, Mutation.position)
.filter(Mutation.is_confirmed == True)
.yield_per(5000)
)
count_mutations(query)
print('Mutations loaded')
collect()
def iterate_known_muts_sources():
for value in tqdm(bdb.values(), total=len(bdb.db)):
for item in map(decode_csv, value):
sources = mutations.get(str(item['protein_id']) + item['alt'] + str(item['pos']))
if sources:
yield sources
counts = defaultdict(int)
fields_ids = [source_id for source_id in sources_map.keys()]
for sources in iterate_known_muts_sources():
for field in fields_ids:
if field in sources:
counts[field] += 1
return {
'Nucleotide mappings': {
sources_map[key]: value
for key, value in counts.items()
}
}
def test_mutation(self):
s = Site(position=13, types={SiteType(name='methylation')})
p = Protein(refseq='NM_007',
id=1,
sites=[s],
sequence='A' * 15,
gene=Gene(name='SomeGene'))
db.session.add(p)
from database import bdb
muts = {13: 14370, 15: 14376}
for aa_pos, dna_pos in muts.items():
muts[aa_pos] = Mutation(protein=p, position=aa_pos, alt='V')
bdb.add_genomic_mut('20',
dna_pos,
'G',
'A',
muts[aa_pos],
is_ptm=True)
query_url = '/chromosome/mutation/{chrom}/{pos}/{ref}/{alt}'
# query as a novel mutation
response = self.client.get(
query_url.format(chrom='chr20', pos=14370, ref='G', alt='A'))
assert response.status_code == 200
assert response.json == [{
'alt':
'V',
'gene':
'SomeGene',
'in_datasets': {},
'pos':
13,
'ptm_impact':
'direct',
'cnt_ptm':
1,
'closest_sites': ['13 A'],
'protein':
'NM_007',
'sites': [{
'kinases': [],
'position': 13,
'residue': 'A',
'kinase_groups': [],
'type': 'methylation'
}],
'ref':
'A'
}]
# well let's look on a known mutation:
m = muts[15]
mc3 = MC3Mutation(mutation=m,
cancer=Cancer(name='Breast invasive carcinoma',
code='BRCA'),
count=1)
esp = ExomeSequencingMutation(mutation=m, maf_all=0.02, maf_aa=0.02)
db.session.add_all([m, mc3, esp])
db.session.commit()
mutation_a15v_query = query_url.format(chrom='chr20',
pos=14376,
ref='G',
alt='A')
response = self.client.get(mutation_a15v_query)
metadata = {
'MC3': {
'Cancers': [{
'Cancer': 'Breast invasive carcinoma',
'Value': 1
}]
},
'ESP6500': {
'MAF': 0.02,
'MAF AA': 0.02,
'MAF EA': None
}
}
assert response.json[0]['in_datasets'] == metadata
expected_values = {'MC3': 1, 'ESP6500': 0.02}
# if user does not want to download data for all datasets he may use:
for source, meta in metadata.items():
response = self.client.get(mutation_a15v_query +
'?filters=Mutation.sources:in:' +
source)
json = response.json[0]
assert json['in_datasets'] == {source: meta}
assert json['value'] == expected_values[source]
response = self.client.get(
mutation_a15v_query +
'?filters=Mutation.sources:in:MC3;Mutation.mc3_cancer_code:in:BRCA'
)
assert response.json
response = self.client.get(
mutation_a15v_query +
'?filters=Mutation.sources:in:ESP6500;Mutation.populations_ESP6500:in:African American'
)
assert response.json
response = self.client.get(
mutation_a15v_query +
'?filters=Mutation.sources:in:ESP6500;Mutation.populations_ESP6500:in:European American'
)
assert not response.json