def _add_variant_protein_variant_assoc_to_graph(self, row):
"""
Generates relationships between variants and protein variants
given a row of data
:param iterable: row of data, see add_variant_info_to_graph()
docstring for expected structure
:return None
"""
gu = GraphUtils(curie_map.get())
geno = Genotype(self.graph)
is_missense = False
is_literal = True
(variant_key, variant_label, amino_acid_variant, amino_acid_position,
transcript_id, transcript_priority, protein_variant_type,
functional_impact, stop_gain_loss, transcript_gene,
protein_variant_source) = row[0:11]
variant_id = self.make_cgd_id('variant{0}'.format(variant_key))
transcript_curie = self._make_transcript_curie(transcript_id)
uniprot_curie = self._make_uniprot_polypeptide_curie(transcript_id)
ncbi_protein_curie = self._make_ncbi_polypeptide_curie(transcript_id)
geno.addGenotype(variant_id, variant_label,
geno.genoparts['sequence_alteration'])
# Make fake amino acid sequence in case we
# can't get a CCDS to Uniprot and/or NCBI Protein mapping
aa_seq_id = self.make_cgd_id('transcript{0}'.format(amino_acid_variant))
# Add Transcript:
geno.addTranscript(variant_id, transcript_curie, transcript_id,
geno.genoparts['transcript'])
# Add polypeptide
if ncbi_protein_curie is not None:
geno.addPolypeptide(ncbi_protein_curie,
self.transcript_xrefs['RefSeq'][transcript_id],
transcript_curie)
aa_seq_id = ncbi_protein_curie
if uniprot_curie is not None:
geno.addPolypeptide(uniprot_curie,
self.transcript_xrefs['UniProt'][transcript_id],
transcript_curie)
# Overrides ncbi_protein_curie,
# but we set them as equal individuals below
aa_seq_id = uniprot_curie
if ncbi_protein_curie is not None and uniprot_curie is not None:
gu.addSameIndividual(self.graph, ncbi_protein_curie, uniprot_curie)
else:
aa_seq_id = self.make_cgd_id('transcript{0}'.format(amino_acid_variant))
if protein_variant_type == 'nonsynonymous - missense' \
or re.search(r'missense', variant_label):
is_missense = True
geno.addGenotype(variant_id, variant_label,
geno.genoparts['missense_variant'])
# Get gene ID from gene map
self._add_variant_gene_relationship(variant_id, transcript_gene)
amino_acid_regex = re.compile(r'^p\.([A-Za-z]{1,3})(\d+)([A-Za-z]{1,3})$')
if is_missense:
match = re.match(amino_acid_regex, amino_acid_variant.rstrip())
else:
match = None
if match is not None:
ref_amino_acid = match.group(1)
position = match.group(2)
altered_amino_acid = match.group(3)
else:
logger.debug("Could not parse amino acid information"
" from {0} variant:"
" {1} type: {2}".format(amino_acid_variant,
variant_label,
protein_variant_type))
# Add amino acid change to model
if is_missense is True and match is not None:
gu.addTriple(self.graph, variant_id,
geno.properties['reference_amino_acid'],
ref_amino_acid, is_literal)
gu.addTriple(self.graph, variant_id,
geno.properties['results_in_amino_acid_change'],
altered_amino_acid, is_literal)
aa_region_id = ":_{0}{1}{2}Region".format(position, position, aa_seq_id)
self._add_feature_with_coords(variant_id, position,
position, aa_seq_id, aa_region_id)
return
