def build_mhc1_alleles(alleles: List[str],
mhc_database: MhcDatabase) -> List[Mhc1]:
isoforms = []
try:
mhc_parser = MhcParser.get_mhc_parser(mhc_database)
# NOTE: during the pandas parsing of empty columns empty lists become a list with one empty string
parsed_alleles = list(
map(mhc_parser.parse_mhc_allele,
filter(lambda a: a != "", alleles)))
for a in parsed_alleles:
ModelValidator.validate_mhc1_gene(a)
# do we need to validate genes anymore? add test creating MhcAllele with bad gene and see what happens
for mhc1_gene in mhc_database.mhc1_genes:
gene_alleles = list(
filter(lambda a: a.gene == mhc1_gene.name, parsed_alleles))
zygosity = MhcFactory._get_zygosity_from_alleles(gene_alleles)
if zygosity == Zygosity.HOMOZYGOUS:
gene_alleles = [
gene_alleles[0]
] # we don't want repeated instances of the same allele
isoforms.append(
Mhc1(name=mhc1_gene,
zygosity=zygosity,
alleles=gene_alleles))
except AssertionError as e:
raise NeofoxDataValidationException(e)
return list(filter(lambda i: i.zygosity != Zygosity.LOSS, isoforms))
def __init__(self, generator, mhc1_alleles, mhc2_alleles, hla_database: HlaDatabase):
Provider.__init__(self, generator)
self.hla_database = hla_database
self.mhc_parser = MhcParser.get_mhc_parser(hla_database)
# gets available alleles from netmhcpan and netmhc2pan
self.available_mhc1_alleles = self.load_mhc1_alleles(mhc1_alleles)
self.available_mhc2_isoforms = self.load_mhc2_isoforms(mhc2_alleles)
# gets available tumor types
self.available_tumor_types = ExpressionAnnotator().cohort_indices.keys()
def setUp(self):
references, self.configuration = integration_test_tools.load_references(organism=ORGANISM_MUS_MUSCULUS)
self.runner = Runner()
self.available_alleles = references.get_available_alleles()
self.test_mhc_one = integration_test_tools.get_h2_one_test(references.get_mhc_database())
self.test_mhc_two = integration_test_tools.get_h2_two_test(references.get_mhc_database())
self.mhc_parser = MhcParser.get_mhc_parser(references.get_mhc_database())
self.proteome_blastp_runner = BlastpRunner(
runner=self.runner, configuration=self.configuration,
database=references.get_proteome_database())
def load_mhc2_alleles(self, available_alleles: List[str]):
mhc_alleles = []
for a in available_alleles:
try:
parsed_allele = MhcParser.get_mhc_parser(
self.hla_database).parse_mhc2_isoform(a)
except AssertionError:
continue
mhc_alleles.append(parsed_allele.name)
return mhc_alleles
def setUp(self):
self.references, self.configuration = integration_test_tools.load_references()
self.runner = Runner()
mhc_parser = MhcParser.get_mhc_parser(self.references.get_mhc_database())
self.mixmhcpred = MixMHCpred(
runner=self.runner, configuration=self.configuration, mhc_parser=mhc_parser
)
self.mixmhc2pred = MixMhc2Pred(
runner=self.runner, configuration=self.configuration, mhc_parser=mhc_parser
)
self.hla_database = self.references.get_mhc_database()
self.test_mhc_one = integration_test_tools.get_hla_one_test(self.hla_database)
self.test_mhc_two = integration_test_tools.get_hla_two_test(self.hla_database)
self.uniprot = Uniprot(self.references.uniprot_pickle)
def build_mhc2_alleles(alleles: List[str],
mhc_database: MhcDatabase) -> List[Mhc2]:
mhc2s = []
try:
mhc_parser = MhcParser.get_mhc_parser(mhc_database)
# NOTE: during the pandas parsing of empty columns empty lists become a list with one empty string
parsed_alleles = list(
map(mhc_parser.parse_mhc_allele,
filter(lambda a: a != "", alleles)))
for a in parsed_alleles:
ModelValidator.validate_mhc2_gene(a)
# do we need to validate genes anymore? add test creating MhcAllele with bad gene and see what happens
for mhc2_isoform_name in mhc_database.mhc2_molecules:
mhc2_isoform_genes = GENES_BY_MOLECULE.get(mhc2_isoform_name)
isoform_alleles = list(
filter(
lambda a: a.gene in
[g.name for g in mhc2_isoform_genes], parsed_alleles))
genes = []
for gene_name in mhc2_isoform_genes:
gene_alleles = list(
filter(lambda a: a.gene == gene_name.name,
isoform_alleles))
zygosity = MhcFactory._get_zygosity_from_alleles(
gene_alleles)
if zygosity == Zygosity.HOMOZYGOUS:
gene_alleles = [
gene_alleles[0]
] # we don't want repeated instances of the same allele
genes.append(
Mhc2Gene(name=gene_name,
zygosity=zygosity,
alleles=gene_alleles))
isoforms = MhcFactory._get_mhc2_isoforms(
mhc2_isoform_name, genes)
mhc2s.append(
Mhc2(name=mhc2_isoform_name,
genes=genes,
isoforms=isoforms))
except AssertionError as e:
raise NeofoxDataValidationException(e)
return list(
filter(
lambda m: all(
map(lambda g: g.zygosity != Zygosity.LOSS, m.genes)),
mhc2s))
def __init__(
self,
references: ReferenceFolder,
configuration: DependenciesConfiguration,
tcell_predictor: TcellPrediction,
self_similarity: SelfSimilarityCalculator,
affinity_threshold =neofox.AFFINITY_THRESHOLD_DEFAULT
):
"""class to annotate neoantigens"""
self.runner = Runner()
self.configuration = configuration
self.proteome_db = references.proteome_db
self.available_alleles = references.get_available_alleles()
self.tcell_predictor = tcell_predictor
self.self_similarity = self_similarity
self.organism = references.organism
# NOTE: this one loads a big file, but it is faster loading it multiple times than passing it around
self.uniprot = Uniprot(references.uniprot_pickle)
# initialise proteome and IEDB BLASTP runners
self.proteome_blastp_runner = BlastpRunner(
runner=self.runner, configuration=configuration,
database=references.get_proteome_database())
self.iedb_blastp_runner = BlastpRunner(
runner=self.runner, configuration=configuration,
database=references.get_iedb_database())
# NOTE: these resources do not read any file thus can be initialised fast
self.dissimilarity_calculator = DissimilarityCalculator(
proteome_blastp_runner=self.proteome_blastp_runner, affinity_threshold=affinity_threshold)
self.neoantigen_fitness_calculator = NeoantigenFitnessCalculator(iedb_blastp_runner=self.iedb_blastp_runner)
self.neoag_calculator = NeoagCalculator(
runner=self.runner, configuration=configuration, affinity_threshold=affinity_threshold
)
self.differential_binding = DifferentialBinding(affinity_threshold=affinity_threshold)
self.priority_score_calculator = PriorityScore()
self.iedb_immunogenicity = IEDBimmunogenicity(affinity_threshold=affinity_threshold)
self.amplitude = Amplitude()
self.hex = Hex(runner=self.runner, configuration=configuration, references=references)
self.mhc_database = references.get_mhc_database()
self.mhc_parser = MhcParser.get_mhc_parser(self.mhc_database)
self.resources_versions = references.get_resources_versions()
def setUp(self):
references, self.configuration = integration_test_tools.load_references(
)
self.runner = Runner()
self.available_alleles_mhc1 = (
references.get_available_alleles().get_available_mhc_i())
self.available_alleles_mhc2 = (
references.get_available_alleles().get_available_mhc_ii())
self.hla_database = references.get_mhc_database()
self.mhc_parser = MhcParser.get_mhc_parser(self.hla_database)
self.test_mhc_one = integration_test_tools.get_hla_one_test(
self.hla_database)
self.test_mhc_two = integration_test_tools.get_hla_two_test(
self.hla_database)
self.uniprot = Uniprot(references.uniprot_pickle)
self.proteome_blastp_runner = BlastpRunner(
runner=self.runner,
configuration=self.configuration,
database=references.get_proteome_database())
def setUp(self) -> None:
self.mhc_parser = MhcParser.get_mhc_parser(FakeHlaDatabase())