def test_mixmhc2pred_allele(self):
mutation = get_mutation(mutated_xmer="TNENLDLQELVEKLEKN", wild_type_xmer="TNENLDLQNLVEKLEKN")
# this is a MHC II genotype which results in no available alleles for MixMHC2pred
MHC_TWO_NEW = MhcFactory.build_mhc2_alleles(
[
"HLA-DRB1*14:54",
"HLA-DRB1*14:54",
"HLA-DQA1*01:04",
"HLA-DQA1*01:04",
"HLA-DQB1*05:03",
"HLA-DQB1*05:03",
"HLA-DPB1*02:01",
"HLA-DPB1*02:01"
],
self.hla_database
)
alleles = self.mixmhc2pred.transform_hla_ii_alleles_for_prediction(MHC_TWO_NEW)
logger.info(alleles)
best_peptide, best_rank, best_allele = self.mixmhc2pred.run(
mutation=mutation, mhc=MHC_TWO_NEW, uniprot=self.uniprot
)
logger.info(best_peptide)
self.assertIsNone(best_peptide)
self.assertIsNone(best_rank)
self.assertIsNone(best_allele)
def test_parse_mhc1_hemizygous_alleles(self):
mhc1s = MhcFactory.build_mhc1_alleles(
["HLA-A*01:01", "HLA-B*07:02", "HLA-C*01:02"], self.hla_database
)
self.assertEqual(3, len(mhc1s))
for mhc1 in mhc1s:
self.assertEqual(Zygosity.HEMIZYGOUS, mhc1.zygosity)
self.assertEqual(1, len(mhc1.alleles))
def get_hla_one_test(hla_database):
return MhcFactory.build_mhc1_alleles([
"HLA-A*24:02",
"HLA-A*02:01",
"HLA-B*15:01",
"HLA-B*44:02",
"HLA-C*07:02",
"HLA-C*05:01",
], hla_database)
def get_h2_one_test(h2_database):
return MhcFactory.build_mhc1_alleles([
"H2Kd",
"H2Kd",
"H2Dd",
"H2Dd",
"H2Ld",
"H2Ld",
], h2_database)
def test_parse_mhc1_non_existing_allele_does_not_fail_mouse(self):
mhc1s = MhcFactory.build_mhc1_alleles(
[
"H2Kd",
"H2Kz", # this one does not exist
"H2Dd",
"H2Ld"
], self.h2_database
)
self.assertEqual(3, len(mhc1s))
def test_parse_mhc2_non_existing_allele_does_not_fail_mouse(self):
mhc2s = MhcFactory.build_mhc2_alleles(
[
"H2Ad",
"H2Az", # this one does not exist
"H2Ed",
"H2Ed",
],
self.h2_database
)
self.assertEqual(2, len(mhc2s))
def test_parse_mhc1_hemizygous_alleles_mouse(self):
mhc1s = MhcFactory.build_mhc1_alleles(
[
"H2Kd",
"H2Dd",
"H2Ld"
], self.h2_database
)
self.assertEqual(3, len(mhc1s))
for mhc1 in mhc1s:
self.assertEqual(Zygosity.HEMIZYGOUS, mhc1.zygosity)
self.assertEqual(1, len(mhc1.alleles))
def test_parse_mhc1_non_existing_allele_does_not_fail(self):
mhc1s = MhcFactory.build_mhc1_alleles(
[
"HLA-A*01:01",
"HLA-A*01:01",
"HLA-B*999:01", # this one does not exist
"HLA-B*07:02",
"HLA-C*01:02",
"HLA-C*01:02",
], self.hla_database
)
self.assertEqual(3, len(mhc1s))
def test_parse_mhc2_non_existing_allele_does_not_fail(self):
mhc2s = MhcFactory.build_mhc2_alleles(
[
"HLA-DRB1*999:01", # this one does not exist
"HLA-DPA1*01:03",
"HLA-DPB1*01:01",
"HLA-DQA1*01:01",
"HLA-DQB1*02:01",
],
self.hla_database
)
self.assertEqual(3, len(mhc2s))
def get_hla_two_test(hla_database):
return MhcFactory.build_mhc2_alleles([
"HLA-DRB1*04:02",
"HLA-DRB1*08:01",
"HLA-DQA1*03:01",
"HLA-DQA1*04:01",
"HLA-DQB1*03:02",
"HLA-DQB1*04:02",
"HLA-DPA1*01:03",
"HLA-DPA1*02:01",
"HLA-DPB1*13:01",
"HLA-DPB1*04:01",
], hla_database)
def patient(self) -> Patient:
patient = None
found = False
while not found:
dr_isoforms = self.random_elements(self.get_hla_ii_alleles_by_gene(Mhc2Name.DR), unique=True, length=2)
dp_isoforms = self.random_elements(self.get_hla_ii_alleles_by_gene(Mhc2Name.DP), unique=True, length=2)
dq_isoforms = self.random_elements(self.get_hla_ii_alleles_by_gene(Mhc2Name.DQ), unique=True, length=2)
# NOTE: for some reason some DP alleles are malformed and cause a validation error, most do not.
# thus I retry until I get a valid combination of HLA alleles, will clarify in another reincarnation
try:
patient = Patient(
identifier=self.generator.unique.uuid4(),
is_rna_available=True,
tumor_type=self.random_elements(self.available_tumor_types, length=1)[0],
# by setting unique=True we enforce that all patients are heterozygous
mhc1=MhcFactory.build_mhc1_alleles(
self.random_elements(self.get_hla_i_alleles_by_gene(Mhc1Name.A), unique=True, length=2) +
self.random_elements(self.get_hla_i_alleles_by_gene(Mhc1Name.B), unique=True, length=2) +
self.random_elements(self.get_hla_i_alleles_by_gene(Mhc1Name.C), unique=True, length=2),
self.hla_database
),
mhc2=MhcFactory.build_mhc2_alleles(
[i.alpha_chain.name for i in dp_isoforms] +
[i.beta_chain.name for i in dp_isoforms] +
[i.alpha_chain.name for i in dq_isoforms] +
[i.beta_chain.name for i in dq_isoforms] +
[i.beta_chain.name for i in dr_isoforms],
self.hla_database
)
)
ModelValidator.validate_patient(patient)
except NeofoxDataValidationException:
continue
found = True
return patient
def test_patient_with_non_existing_allele_does_not_crash(self):
""""""
neoantigens, patients, patient_id = self._get_test_data()
for p in patients:
# sets one MHC I allele to a non existing allele
p.mhc1[0].alleles[0] = MhcFactory.build_mhc1_alleles(
["HLA-A*99:99"], mhc_database=self.hla_database)[0].alleles[0]
neofox = NeoFox(
neoantigens=neoantigens,
patient_id=patient_id,
patients=patients,
num_cpus=1,
)
neofox.get_annotations()
def test_mixmhcpred_not_supported_allele(self):
"""
this is a combination of neoepitope and HLA alleles from Balachandran
"""
mutation = get_mutation(mutated_xmer="SIYGGLVLI", wild_type_xmer="PIYGGLVLI")
best_peptide, best_rank, best_allele, best_score = self.mixmhcpred.run(
mutation=mutation,
mhc=MhcFactory.build_mhc1_alleles(["A02:01", "B44:02", "C05:17", "C05:01"], self.hla_database),
uniprot=self.uniprot
)
self.assertEqual('SIYGGLVLI', best_peptide)
self.assertEqual(0.15829400000000002, best_score)
self.assertEqual(1, best_rank)
self.assertEqual('HLA-A*02:01', best_allele)
def test_parse_mhc2_hemizygous_alleles_mouse(self):
mhc2s = MhcFactory.build_mhc2_alleles(
[
"H2Ad",
"H2Ed",
],
self.h2_database
)
self.assertEqual(2, len(mhc2s))
for mhc2 in mhc2s:
self.assertEqual(1, len(mhc2.genes))
for gene in mhc2.genes:
self.assertEqual(1, len(gene.alleles))
self.assertEqual(Zygosity.HEMIZYGOUS, gene.zygosity)
self.assertEqual(1 if mhc2.name == Mhc2Name.DR else 1, len(mhc2.isoforms))
self._assert_isoforms(mhc2)
def test_parse_mhc2_homozygous_alleles_mouse(self):
mhc2s = MhcFactory.build_mhc2_alleles(
[
"H2Ad",
"H2Ad",
"H2Ed",
"H2Ed"
],
self.h2_database
)
self.assertEqual(2, len(mhc2s))
for mhc2 in mhc2s:
self.assertEqual(1, len(mhc2.genes))
for gene in mhc2.genes:
self.assertEqual(1, len(gene.alleles))
self.assertEqual(Zygosity.HOMOZYGOUS, gene.zygosity)
self.assertEqual(1, len(mhc2.isoforms))
self._assert_isoforms(mhc2)
def test_parse_mhc2_hemizygous_alleles(self):
mhc2s = MhcFactory.build_mhc2_alleles(
[
"HLA-DRB1*01:01",
"HLA-DPA1*01:03",
"HLA-DPB1*02:01",
"HLA-DQA1*01:01",
"HLA-DQB1*02:01",
],
self.hla_database
)
self.assertEqual(3, len(mhc2s))
for mhc2 in mhc2s:
self.assertEqual(1 if mhc2.name == Mhc2Name.DR else 2, len(mhc2.genes))
for gene in mhc2.genes:
self.assertEqual(1, len(gene.alleles))
self.assertEqual(Zygosity.HEMIZYGOUS, gene.zygosity)
self.assertEqual(1 if mhc2.name == Mhc2Name.DR else 1, len(mhc2.isoforms))
self._assert_isoforms(mhc2)
def test_parse_mhc2_hetero_and_homozygous_alleles(self):
mhc2s = MhcFactory.build_mhc2_alleles(
[
"HLA-DRB1*01:01",
"HLA-DRB1*01:01",
"HLA-DPA1*01:03",
"HLA-DPA1*01:03",
"HLA-DPB1*02:01",
"HLA-DPB1*02:02",
"HLA-DQA1*01:01",
"HLA-DQA1*01:01",
"HLA-DQB1*02:01",
"HLA-DQB1*02:02",
],
self.hla_database
)
self.assertEqual(3, len(mhc2s))
for mhc2 in mhc2s:
self.assertEqual(1 if mhc2.name == Mhc2Name.DR else 2, len(mhc2.genes))
self.assertEqual(1 if mhc2.name == Mhc2Name.DR else 2, len(mhc2.isoforms))
def test_neoantigen_no_wt_failing(self):
patient_identifier = "12345"
neoantigen = Neoantigen(
mutation=Mutation(mutated_xmer="SPSFPLEPDDEVFTAIAKAMEEMVEDS"),
patient_identifier=patient_identifier)
patient = Patient(
identifier=patient_identifier,
mhc1=MhcFactory.build_mhc1_alleles(
[
"HLA-A*02:24", "HLA-A*36:04", "HLA-B*58:25",
"HLA-B*35:102", "HLA-C*02:30", "HLA-C*07:139"
],
mhc_database=self.references.get_mhc_database()),
)
annotations = NeoFox(
neoantigens=[neoantigen],
patients=[patient],
num_cpus=1,
).get_annotations()
# it does not crash even though there are no best 9mers
self.assertIsNotNone(annotations)
def test_neoantigen_failing(self):
patient_identifier = "12345"
neoantigen = Neoantigen(mutation=Mutation(
wild_type_xmer="ARPDMFCLFHGKRYFPGESWHPYLEPQ",
mutated_xmer="ARPDMFCLFHGKRHFPGESWHPYLEPQ"),
patient_identifier=patient_identifier)
patient = Patient(
identifier=patient_identifier,
mhc1=MhcFactory.build_mhc1_alleles(
[
"HLA-A*03:01", "HLA-A*29:02", "HLA-B*07:02", "HLA-B*44:03",
"HLA-C*07:02", "HLA-C*16:01"
],
mhc_database=self.references.get_mhc_database()),
)
annotations = NeoFox(
neoantigens=[neoantigen],
patients=[patient],
num_cpus=1,
).get_annotations()
# it does not crash even though there are no best 9mers
self.assertIsNotNone(annotations)
def test_parse_mhc1_loss_alleles(self):
mhc1s = MhcFactory.build_mhc1_alleles([], self.hla_database)
self.assertEqual(0, len(mhc1s))
def get_h2_two_test(h2_database):
return MhcFactory.build_mhc2_alleles(["H2Ad", "H2Ad", "H2Ed", "H2Ed"],
h2_database)
def test_phbr1(self):
best_multiple = BestAndMultipleBinder(
runner=self.runner,
configuration=self.configuration,
mhc_parser=self.mhc_parser,
blastp_runner=self.proteome_blastp_runner)
netmhcpan = NetMhcPanPredictor(
runner=self.runner,
configuration=self.configuration,
mhc_parser=self.mhc_parser,
blastp_runner=self.proteome_blastp_runner)
mutation = get_mutation(
mutated_xmer="DEVLGEPSQDILVTDQTRLEATISPET",
wild_type_xmer="DEVLGEPSQDILVIDQTRLEATISPET",
)
# all alleles = heterozygous
predictions = netmhcpan.mhc_prediction(self.test_mhc_one,
self.available_alleles_mhc1,
mutation.mutated_xmer)
predicted_neoepitopes = netmhcpan.remove_peptides_in_proteome(
predictions=predictions, uniprot=self.uniprot)
best_epitopes_per_allele = (
BestAndMultipleBinder.extract_best_epitope_per_alelle(
predicted_neoepitopes, self.test_mhc_one))
phbr_i = best_multiple.calculate_phbr_i(best_epitopes_per_allele,
self.test_mhc_one)
self.assertIsNotNone(phbr_i)
self.assertAlmostEqual(1.359324592015038, phbr_i)
# one homozygous allele present
mhc_alleles = MhcFactory.build_mhc1_alleles([
"HLA-A*24:02",
"HLA-A*02:01",
"HLA-B*15:01",
"HLA-B*44:02",
"HLA-C*05:01",
"HLA-C*05:01",
], self.hla_database)
predictions = netmhcpan.mhc_prediction(self.test_mhc_one,
self.available_alleles_mhc1,
mutation.mutated_xmer)
predicted_neoepitopes = netmhcpan.remove_peptides_in_proteome(
predictions=predictions, uniprot=self.uniprot)
best_epitopes_per_allele = (
BestAndMultipleBinder.extract_best_epitope_per_alelle(
predicted_neoepitopes, mhc_alleles))
phbr_i = best_multiple.calculate_phbr_i(best_epitopes_per_allele,
mhc_alleles)
self.assertIsNotNone(phbr_i)
self.assertAlmostEqual(1.0036998409510969, phbr_i)
# mo info for one allele
mhc_alleles = MhcFactory.build_mhc1_alleles([
"HLA-A*24:02", "HLA-A*02:01", "HLA-B*15:01", "HLA-B*44:02",
"HLA-C*05:01"
], self.hla_database)
predictions = netmhcpan.mhc_prediction(self.test_mhc_one,
self.available_alleles_mhc1,
mutation.mutated_xmer)
predicted_neoepitopes = netmhcpan.remove_peptides_in_proteome(
predictions=predictions, uniprot=self.uniprot)
best_epitopes_per_allele = (
BestAndMultipleBinder.extract_best_epitope_per_alelle(
predicted_neoepitopes, mhc_alleles))
phbr_i = best_multiple.calculate_phbr_i(best_epitopes_per_allele,
mhc_alleles)
self.assertIsNone(phbr_i)
def test_phbr2(self):
best_multiple = BestAndMultipleBinderMhcII(
runner=self.runner,
configuration=self.configuration,
mhc_parser=self.mhc_parser,
blastp_runner=self.proteome_blastp_runner)
netmhc2pan = NetMhcIIPanPredictor(
runner=self.runner,
configuration=self.configuration,
mhc_parser=self.mhc_parser,
blastp_runner=self.proteome_blastp_runner)
mutation = get_mutation(
mutated_xmer="DEVLGEPSQDILVTDQTRLEATISPET",
wild_type_xmer="DEVLGEPSQDILVIDQTRLEATISPET",
)
# all alleles = heterozygous
allele_combinations = netmhc2pan.generate_mhc2_alelle_combinations(
self.test_mhc_two)
patient_mhc2_isoforms = best_multiple._get_only_available_combinations(
allele_combinations, self.available_alleles_mhc2)
predictions = netmhc2pan.mhc2_prediction(patient_mhc2_isoforms,
mutation.mutated_xmer)
filtered_predictions = netmhc2pan.remove_peptides_in_proteome(
predictions=predictions, uniprot=self.uniprot)
logger.info(filtered_predictions)
logger.info(self.test_mhc_two)
best_predicted_epitopes_per_alelle = (
best_multiple.extract_best_epitope_per_mhc2_alelle(
predictions=filtered_predictions,
mhc_isoforms=self.test_mhc_two))
phbr_ii = best_multiple.calculate_phbr_ii(
best_predicted_epitopes_per_alelle)
self.assertIsNotNone(phbr_ii)
self.assertAlmostEqual(8.895757526065129, phbr_ii)
# mo info for one allele
mhc2_alleles = MhcFactory.build_mhc2_alleles([
"HLA-DRB1*04:02",
"HLA-DRB1*08:01",
"HLA-DQA1*03:01",
"HLA-DQA1*04:01",
"HLA-DQB1*03:02",
"HLA-DQB1*04:02",
"HLA-DPA1*01:03",
"HLA-DPA1*02:01",
"HLA-DPB1*13:01",
"HLA-DPB1*13:01",
], self.hla_database)
allele_combinations = netmhc2pan.generate_mhc2_alelle_combinations(
mhc2_alleles)
patient_mhc2_isoforms = best_multiple._get_only_available_combinations(
allele_combinations, self.available_alleles_mhc2)
predictions = netmhc2pan.mhc2_prediction(patient_mhc2_isoforms,
mutation.mutated_xmer)
filtered_predictions = netmhc2pan.remove_peptides_in_proteome(
predictions=predictions, uniprot=self.uniprot)
best_predicted_epitopes_per_alelle = (
best_multiple.extract_best_epitope_per_mhc2_alelle(
filtered_predictions, mhc2_alleles))
logger.info(best_predicted_epitopes_per_alelle)
logger.info(len(best_predicted_epitopes_per_alelle))
phbr_ii = best_multiple.calculate_phbr_ii(
best_predicted_epitopes_per_alelle)
self.assertIsNone(phbr_ii)
# one allele present
mhc2_alleles = MhcFactory.build_mhc2_alleles([
"HLA-DRB1*04:02",
"HLA-DRB1*08:01",
"HLA-DQA1*03:01",
"HLA-DQA1*04:01",
"HLA-DQB1*03:02",
"HLA-DQB1*04:02",
"HLA-DPA1*01:03",
"HLA-DPA1*02:01",
"HLA-DPB1*13:01",
], self.hla_database)
allele_combinations = netmhc2pan.generate_mhc2_alelle_combinations(
mhc2_alleles)
patient_mhc2_isoforms = best_multiple._get_only_available_combinations(
allele_combinations, self.available_alleles_mhc2)
predictions = netmhc2pan.mhc2_prediction(patient_mhc2_isoforms,
mutation.mutated_xmer)
filtered_predictions = netmhc2pan.remove_peptides_in_proteome(
predictions=predictions, uniprot=self.uniprot)
best_predicted_epitopes_per_alelle = (
best_multiple.extract_best_epitope_per_mhc2_alelle(
filtered_predictions, mhc2_alleles))
logger.info(best_predicted_epitopes_per_alelle)
logger.info(len(best_predicted_epitopes_per_alelle))
phbr_ii = best_multiple.calculate_phbr_ii(
best_predicted_epitopes_per_alelle)
self.assertIsNone(phbr_ii)
def test_parse_mhc2_loss_mouse(self):
mhc2s = MhcFactory.build_mhc2_alleles([], self.h2_database)
self.assertEqual(0, len(mhc2s))