def predict(self, peptides_df):
"""
Determine whether 9-mer peptide is immunogenic by checking
1) that the epitope binds strongly to a particular MHC allele
2) the "core" of the peptide (positions 3-8) don't overlap with any
other peptides in the self/thymic MHC ligand sets of that HLA allele
Returns DataFrame with two extra columns:
- ThymicDeletion: Was this epitope deleted during thymic selection
(and thus can't be recognize by T-cells)?
- Immunogenic: Is this epitope a sufficiently strong binder that
wasn't deleted during thymic selection?
"""
thymic_peptide_sets = self.peptide_sets.values()
# assume a peptide is non-immunogenic unless not in thymic sets
# We do this in case some alleles are missing, resulting in all
# their associated ligands being considered non-immunogenic
peptides_df[THYMIC_DELETION_FIELD_NAME] = True
for i in xrange(len(peptides_df)):
row = peptides_df.ix[i]
peptide = row.Epitope
allele = compact_hla_allele_name(row.Allele)
if allele in self.peptide_sets:
# positions in the epitope are indexed starting from 1 to
# match immunology nomenclature
substring = \
peptide[self.first_position - 1 : self.last_position]
peptides_df[THYMIC_DELETION_FIELD_NAME].ix[i] = \
substring in self.peptide_sets[allele]
peptides_df["Immunogenic"] = \
~peptides_df[THYMIC_DELETION_FIELD_NAME] & \
(peptides_df[IC50_FIELD_NAME] <= self.binding_threshold)
return peptides_df
def predict(self, peptides_df):
"""
Determine whether 9-mer peptide is immunogenic by checking
1) that the epitope binds strongly to a particular MHC allele
2) the "core" of the peptide (positions 3-8) don't overlap with any
other peptides in the self/thymic MHC ligand sets of that HLA allele
Returns DataFrame with two extra columns:
- ThymicDeletion: Was this epitope deleted during thymic selection
(and thus can't be recognize by T-cells)?
- Immunogenic: Is this epitope a sufficiently strong binder that
wasn't deleted during thymic selection?
"""
thymic_peptide_sets = self.peptide_sets.values()
# assume a peptide is non-immunogenic unless not in thymic sets
# We do this in case some alleles are missing, resulting in all
# their associated ligands being considered non-immunogenic
peptides_df[THYMIC_DELETION_FIELD_NAME] = True
for i in xrange(len(peptides_df)):
row = peptides_df.ix[i]
peptide = row.Epitope
allele = compact_hla_allele_name(row.Allele)
if allele in self.peptide_sets:
# positions in the epitope are indexed starting from 1 to
# match immunology nomenclature
substring = \
peptide[self.first_position - 1 : self.last_position]
peptides_df[THYMIC_DELETION_FIELD_NAME].ix[i] = \
substring in self.peptide_sets[allele]
peptides_df["Immunogenic"] = \
~peptides_df[THYMIC_DELETION_FIELD_NAME] & \
(peptides_df[IC50_FIELD_NAME] <= self.binding_threshold)
return peptides_df
def __init__(
self,
alleles,
data_path = DEFAULT_PEPTIDE_DIR,
binding_threshold = 500,
first_position = 3,
last_position = 8):
"""
Parameters
--------
alleles : list of strings
data_path : str, optional
first_position : int, optional
Start position for extracting substring of
query peptide (indexed starting from 1)
last_position : int, optional
Last position for extracting substring of
query peptide (indexed starting from 1)
"""
self.binding_threshold = binding_threshold
self.first_position = first_position
self.last_position = last_position
self.alleles = {
compact_hla_allele_name(allele) for allele in alleles
}
self.data_path = data_path
assert exists(self.data_path), \
"Directory with thymic peptides (%s) does not exist" % \
self.data_path
available_alleles = listdir(self.data_path)
mappings_file_path = join(self.data_path, 'mappings')
if exists(mappings_file_path):
self.allele_mappings = \
_load_allele_mapping_dict(mappings_file_path)
else:
self.allele_mappings = \
dict(zip(available_alleles, available_alleles))
self.peptide_sets = {}
for allele in self.alleles:
if allele not in self.allele_mappings:
logging.warn(
"No MHC peptide set available for HLA allele %s", allele)
continue
else:
logging.info(
"Loading thymic MHC peptide set for HLA allele %s", allele)
filename = self.allele_mappings[allele]
assert filename in available_alleles, \
"No MHC peptide set available for HLA allele %s (file = %s)" % \
(allele,filename)
with open(join(self.data_path, filename), 'r') as f:
peptide_set = {l for l in f.read().split("\n") if len(l) > 0}
self.peptide_sets[allele] = peptide_set
def __init__(self,
alleles,
data_path=DEFAULT_PEPTIDE_DIR,
binding_threshold=500,
first_position=3,
last_position=8):
"""
Parameters
--------
alleles : list of strings
data_path : str, optional
first_position : int, optional
Start position for extracting substring of
query peptide (indexed starting from 1)
last_position : int, optional
Last position for extracting substring of
query peptide (indexed starting from 1)
"""
self.binding_threshold = binding_threshold
self.first_position = first_position
self.last_position = last_position
self.alleles = {compact_hla_allele_name(allele) for allele in alleles}
self.data_path = data_path
assert exists(self.data_path), \
"Directory with thymic peptides (%s) does not exist" % \
self.data_path
available_alleles = listdir(self.data_path)
mappings_file_path = join(self.data_path, 'mappings')
if exists(mappings_file_path):
self.allele_mappings = \
_load_allele_mapping_dict(mappings_file_path)
else:
self.allele_mappings = \
dict(zip(available_alleles, available_alleles))
self.peptide_sets = {}
for allele in self.alleles:
if allele not in self.allele_mappings:
logging.warn("No MHC peptide set available for HLA allele %s",
allele)
continue
else:
logging.info(
"Loading thymic MHC peptide set for HLA allele %s", allele)
filename = self.allele_mappings[allele]
assert filename in available_alleles, \
"No MHC peptide set available for HLA allele %s (file = %s)" % \
(allele,filename)
with open(join(self.data_path, filename), 'r') as f:
peptide_set = {l for l in f.read().split("\n") if len(l) > 0}
self.peptide_sets[allele] = peptide_set