def create_test_kinase(name, refseq):
interactor = Kinase(name=name)
kinase_gene = Gene(name='Gene of ' + interactor.name)
kinase_protein = Protein(refseq=refseq, gene=kinase_gene)
interactor.protein = kinase_protein
return interactor
def get_or_create_kinases(
chosen_kinases_names, known_kinases,
known_kinase_groups) -> [Set[Kinase], Set[KinaseGroup]]:
"""Create a subset of known kinases and known kinase groups based on given
list of kinases names ('chosen_kinases_names'). If no kinase or kinase group
of given name is known, it will be created.
Returns a tuple of sets:
kinases, groups
"""
kinases, groups = set(), set()
for name in set(chosen_kinases_names):
# handle kinases group
if name.endswith('_GROUP'):
name = name[:-6]
key = name.lower()
if key not in known_kinase_groups:
known_kinase_groups[key] = KinaseGroup(name=name)
groups.add(known_kinase_groups[key])
# if it's not a group, it surely is a kinase:
else:
key = name.lower()
if key not in known_kinases:
known_kinases[key] = Kinase(
name=name, protein=get_preferred_gene_isoform(name))
kinases.add(known_kinases[key])
return kinases, groups
def parser(line):
# note that the subfamily is often absent
group, family, subfamily = line[2:5]
# the 'gene.clean' [6] fits better to the names
# of kinases used in all other data files
kinase_name = line[6]
# 'group.clean' is not atomic and is redundant with respect to
# family and subfamily. This check assures that in case of a change
# the maintainer would be able to spot the inconsistency easily
clean = family + '_' + subfamily if subfamily else family
assert line[8] == clean
if kinase_name.lower() not in known_kinases:
kinase = Kinase(name=kinase_name,
protein=get_preferred_gene_isoform(kinase_name))
known_kinases[kinase_name.lower()] = kinase
# the 'family' corresponds to 'group' in the all other files
if family.lower() not in known_groups:
group = KinaseGroup(name=family)
known_groups[family.lower()] = group
new_groups.append(group)
known_groups[family.lower()].kinases.append(
known_kinases[kinase_name.lower()])
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 test_train_model(self):
phosphorylation = SiteType(name='phosphorylation')
# non-phosphorylated serine residues are needed to generate negative sites
p = Protein(refseq='NM_007',
sequence='--------SLPA-----------SVIT-------')
g = Gene(isoforms=[p], preferred_isoform=p)
db.session.add(g)
# phosphorylated, with sites
p = Protein(refseq='NM_001',
sequence='--------SPAK-----------SPAR-------')
g = Gene(isoforms=[p], preferred_isoform=p)
db.session.add(g)
k = Kinase(name='CDK1', is_involved_in={phosphorylation})
for pos in [9, 24]:
s = Site(position=pos,
types={phosphorylation},
residue='S',
protein=p,
kinases={k})
db.session.add(s)
db.session.commit()
with TemporaryDirectory() as temp_dir:
model = train_model(phosphorylation,
sequences_dir=temp_dir,
sampling_n=2,
threshold=2)
# the model should have one set of params - for CDK1 kinase
assert len(model) == 1
cdk_params = model.rx2('CDK1')
pwm = cdk_params.rx2('pwm')
# and the position-specific weight matrix should be created
assert pwm
# the very detailed testing should be performed by rMIMP,
# but why not test the basics?
weights_of_central_aa = {
aa: value
for aa, value in zip(pwm.rownames, pwm.rx(True, 8))
}
assert weights_of_central_aa['S'] == max(
weights_of_central_aa.values())
def test_interactions(self):
from models import Protein, Site, Kinase, KinaseGroup
p1 = Protein(sites=[
Site(),
Site(kinases=[Kinase()], kinase_groups=[KinaseGroup()])
])
db.session.add(p1)
p2 = Protein(sites=[Site(kinases=[Kinase()])])
db.session.add(p2)
u_all_interactions = 0
u_kinases_covered = set()
u_kinase_groups_covered = set()
u_proteins_covered = set()
for protein in models.Protein.query.all():
for site in protein.sites:
kinases = site.kinases
kinase_groups = site.kinase_groups
u_all_interactions += len(kinases) + len(kinase_groups)
u_kinases_covered.update(kinases)
u_kinase_groups_covered.update(kinase_groups)
if kinases or kinase_groups:
u_proteins_covered.add(protein)
from stats import Statistics
statistics = Statistics()
all_interactions = statistics.interactions()
kinases_covered = statistics.kinases_covered()
kinase_groups_covered = statistics.kinase_groups_covered()
proteins_covered = statistics.proteins_covered()
assert all_interactions == u_all_interactions
assert kinases_covered == len(u_kinases_covered)
assert kinase_groups_covered == len(u_kinase_groups_covered)
assert proteins_covered == len(u_proteins_covered)
def parser(line):
kinase_name, gene_name = line
protein = get_preferred_gene_isoform(gene_name)
if not protein:
print('No isoform for %s kinase mapped to %s gene!' %
(kinase_name, gene_name))
return
if kinase_name in known_kinases:
kinase = known_kinases[kinase_name]
if kinase.protein and kinase.protein != protein:
print('Overriding kinase-protein association for '
'%s kinase. Old isoform: %s; new isoform: %s.' %
(kinase_name, kinase.protein.refseq, protein.refseq))
kinase.protein = protein
else:
new_kinases.append(Kinase(name=kinase_name, protein=protein))
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 test_classification(self):
"""Following assertion about data file holds:
- 'family' fits better to our 'group' than any other column
- 'gene.clean', not 'Kinase' is being used as kinase name as it fits much better.
"""
existing_kinases = {
name: Kinase(name=name)
for name in ('AKT1', 'Akt2', 'CIT')
}
existing_groups = {name: KinaseGroup(name=name) for name in ('Akt', )}
def add_to_session():
db.session.add_all(existing_kinases.values())
db.session.add_all(existing_groups.values())
filename = make_named_temp_file(raw_gene_list)
add_to_session()
with self.app.app_context():
new_groups = load_kinase_classification(filename)
assert len(new_groups) == 1
new_group = new_groups[0]
assert new_group.name == 'DMPK'
add_to_session()
assert len(new_group.kinases) == 2
assert existing_kinases['CIT'] in new_group.kinases
old_group = existing_groups['Akt']
assert len(old_group.kinases) == 3
assert existing_kinases['AKT1'] in old_group.kinases
assert existing_kinases['Akt2'] in old_group.kinases