def src_gen(self):
retval = []
for obj in self.graph.objects(
predicate=self.PDBo.has_entity_src_genCategory):
qres = rdf.RDFDataSourceBase(url=str(obj), cache=self.cache).query(
"""prefix PDBo: <http://rdf.wwpdb.org/schema/pdbx-v40.owl#>
select ?entity_id ?scientific_name ?strain ?gene where
{{
?s PDBo:has_entity_src_gen ?entity .
?entity
PDBo:entity_src_gen.entity_id ?entity_id ;
PDBo:entity_src_gen.pdbx_gene_src_scientific_name ?scientific_name ;
PDBo:entity_src_gen.gene_src_strain ?strain ;
PDBo:entity_src_gen.pdbx_gene_src_gene ?gene .
}}
""")
for row in qres:
entity_id, scientific_name, strain, gene = [
str(elem) for elem in row
]
retval.append(
dict(entity_id=entity_id,
scientific_name=scientific_name,
strain=strain,
gene=gene))
return retval
def get_type(self):
qres = self.query(
"""prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>
prefix skos: <http://www.w3.org/2004/02/skos/core#>
select ?type where
{{
?s
rdf:type ?type ;
skos:prefLabel ?label .
}}
""")
uri = [str(row[0]) for row in qres]
assert len(uri) == 1
label = [
str(obj) for obj in rdf.RDFDataSourceBase(uri[0]).graph.objects(
subject=URIRef(uri[0]), predicate=RDFS.label)
]
assert len(label) == 1
return label[0]
def sequence_annotation(
self, uri=UniProtDataSourceBase.UNIPROT.Sequence_Annotation):
# http://www.uniprot.org/core/
# http://www.uniprot.org/help/sequence_annotation
qres = rdf.RDFDataSourceBase(str(self.UNIPROT)).query(
"""prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#>
prefix uniprot: <http://purl.uniprot.org/core/>
select ?s ?label ?comment ?see_also where
{{
{{
{{
?s
rdfs:subClassOf+ <{uri}> ;
rdfs:label ?label .
}}
union
{{
?s rdfs:label ?label .
filter( ?s = <{uri}> ).
}}
optional {{ ?s rdfs:comment ?comment }}
optional {{ ?s rdfs:seeAlso ?see_also . }}
}}
}}
""".format(uri=str(uri)))
names = {}
for row in qres:
name, label = str(row[0]), str(row[1])
value = dict(name=name, label=label)
if row[2] is not None:
value['comment'] = str(row[2])
if row[3] is not None:
value['see_also'] = str(row[3])
names[name] = value
qres = self.query(
"""prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>
prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#>
prefix faldo: <http://biohackathon.org/resource/faldo#>
prefix uniprot: <http://purl.uniprot.org/core/>
select ?type ?begin ?end ?s ?comment ?substitution where
{{
?s
rdf:type ?type ;
uniprot:range ?range .
optional {{ ?s rdfs:comment ?comment }} .
optional {{ ?s uniprot:substitution ?substitution }} .
filter( ?type in ({}) ) .
?range faldo:begin ?begin_ ; faldo:end ?end_ .
?begin_ faldo:position ?begin .
?end_ faldo:position ?end .
}}
""".format(', '.join('<{}>'.format(name)
for name in names.keys())))
retval = []
for row in qres:
name, begin, end = str(row[0]), int(row[1]), int(row[2])
value = dict(begin=begin, end=end, type=names[name])
about = str(row[3])
citation = self.citation(about)
if len(citation) > 0:
value['citation'] = citation
if row[4] is not None:
value['comment'] = str(row[4])
if row[5] is not None:
value['substitution'] = str(row[5])
retval.append(value)
return retval
