@prefix ex: <https://sparql.uniprot.org/.well-known/sparql-examples/> .
@prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> .
@prefix schema: <https://schema.org/> .
@prefix sh: <http://www.w3.org/ns/shacl#> .

ex:163_canonical_isoform_does_not_need_to_be_dash_1 a sh:SPARQLExecutable,
        sh:SPARQLSelectExecutable ;
    rdfs:comment "Find reviewed proteins whose canonical isoform does not end in -1, ordered by the highest isoform number used as the canonical sequence. UniProt's up:sequence points to whichever isoform was chosen as canonical, which is not always isoform -1"@en ;
    sh:prefixes _:sparql_examples_prefixes ;
    sh:select """PREFIX up: <http://purl.uniprot.org/core/>
PREFIX uniprotkb: <http://purl.uniprot.org/uniprot/>
PREFIX xsd: <http://www.w3.org/2001/XMLSchema#>

SELECT
  ?protein
  (?sequence AS ?canonicalIsoform)
  ?isoformCount
WHERE
{
  GRAPH <http://sparql.uniprot.org/uniprot> {
    ?protein up:reviewed true .
    ?protein up:sequence ?sequence .
    ?sequence a up:Simple_Sequence .
    OPTIONAL {
      ?sequence a up:External_Sequence .
      BIND(true AS ?isExternalSequence)
      BIND(SUBSTR(STR(?protein), STRLEN(STR(uniprotkb:))) AS ?proteinAc)
      FILTER(CONTAINS(STR(?sequence), ?proteinAc))
    }
    FILTER(?isExternalSequence || !BOUND(?isExternalSequence))
  }
  BIND(xsd:int(STRAFTER(STR(?sequence), "-")) AS ?isoformCount)
} ORDER BY DESC(?isoformCount)""" ;
    schema:keywords "isoform" , "canonical sequence" ;
    schema:target <https://sparql.uniprot.org/sparql/> .