Querying the graph

A cookbook of example SPARQL from queries/examples/*.rq — domain questions, not the pass/fail checks in Build and validation. Run with make query-examples; each was re-run against the live build (see the Overview's last verified build table for graph size).

Datasets tagged with a specific ontology-resolved tumor type (breast carcinoma, NCIT:C4872), with their DOI — answered via the resolved term rather than a free-text guess at how "breast cancer" is spelled in the source data (174 rows):

PREFIX cckp: <https://w3id.org/mc2-center/cckp-portal/>

SELECT ?dataset ?name ?doi WHERE {
  ?dataset a cckp:Dataset ;
           cckp:tumorTypeTerm <http://purl.obolibrary.org/obo/NCIT_C4872> ;
           cckp:datasetName ?name .
  OPTIONAL { ?dataset cckp:doi ?doi }
}

Every Publication linked to a Dataset, with the Publication's resolved PubMed IRI — a real resolvable link instead of a bare numeric PubMed ID (394 rows):

PREFIX cckp: <https://w3id.org/mc2-center/cckp-portal/>

SELECT ?pub ?title ?pubmedIri ?dataset WHERE {
  ?pub a cckp:Publication ;
       cckp:datasetRef ?dataset ;
       cckp:publicationTitle ?title .
  OPTIONAL { ?pub cckp:pubMedIdIri ?pubmedIri }
}

Per-SCDM-Program output rollup, via the cckp:consortiumRef federation edges link_scdm.py mints — a cross-silo question the base portal tables alone can't answer (26 rows, one per Program × class with at least one row):

PREFIX cckp: <https://w3id.org/mc2-center/cckp-portal/>
PREFIX sagecdm: <https://sage-bionetworks.github.io/SageCommonDataModel/>

SELECT ?programName ?type (COUNT(?s) AS ?n) WHERE {
  ?program a sagecdm:Program ;
           sagecdm:name ?programName .
  ?s cckp:consortiumRef ?program ;
     a ?type .
  VALUES ?type { cckp:Dataset cckp:Publication cckp:Tool }
} GROUP BY ?programName ?type ORDER BY ?programName ?type

A performance note: rdflib's engine handles several independent OPTIONAL blocks badly at this graph's size. The single-join-plus- VALUES form above runs in well under a second.