SPARQL Verification Queries
Introduction
The MSE-KG validation pipeline includes a set of SPARQL verification queries that act as automated quality gates. Each query is designed to detect a specific class of data-quality or ontology-conformance issue before new triples are published to the production knowledge graph.
The principle is simple: every verification query is written so that it returns zero rows when the graph is valid. Any row returned by a query signals a concrete violation that must be fixed before the data can proceed through the pipeline.
How verification fits into the pipeline
These queries run after RDF generation and SHACL shape validation. They complement shape-based validation by catching cross-graph consistency issues (e.g., undeclared classes or properties) and domain-specific constraints (e.g., ORCID representation rules) that are difficult to express in SHACL alone.
Summary
| # | File | What it checks |
|---|---|---|
| 1 | verify1.sparql |
ORCIDs must be IRIs, never plain literals |
| 2 | verify2.sparql |
Forbidden BFO temporal predicates must not appear |
| 3 | verify3.sparql |
Every rdf:type class must be declared in the ontology |
| 4 | verify4.sparql |
Every predicate must be a declared property in the ontology |
| 5 | verify5.sparql |
No two textual entities may share the same value |
| 6 | reasoner_disjoint_violation.sparql |
No individual may instantiate two disjoint classes |
Query Details
1 -- ORCID Must Be an IRI, Not a Literal
What it checks:
Detects triples where an ORCID identifier (https://orcid.org/...) is used as a
plain literal value instead of as an IRI resource, but only for subjects within the
MSE-KG namespace (https://purls.helmholtz-metadaten.de/msekg).
Why it matters: ORCIDs are globally unique researcher identifiers and must be represented as IRIs so that they can be dereferenced, linked across datasets, and matched by standard federated queries. Storing them as string literals breaks interoperability and prevents proper entity resolution.
Expected result
0 rows. Every row returned represents a triple where an ORCID that should be an IRI is incorrectly stored as a literal.
##### OCRID must be IRI, not literal
#
# we don't what ORCIDs as literals used together with our resources
#
# OCDIS always should be IRIs
#
# Remark: some other ontologies or KGs are using ORCIDS as literals, we do not cover theses cases.
# That's why we filter ?s to begin with our namespace.
#
SELECT ?s ?p ?o "Literals should not contain sole ORCID (OCRID must be IRI, not literal)" WHERE {
?s ?p ?o .
FILTER regex(str(?s),"^https://purls.helmholtz-metadaten.de/msekg") .
FILTER isLiteral(?o) .
FILTER regex(str(?o), "^https://orcid.org/\\d{4}-\\d{4}-\\d{4}-\\d{3}[0-9X]{1}$")
}
2 -- Forbidden BFO Temporal Predicates
What it checks:
Finds any BFO (Basic Formal Ontology) predicates used in the graph that are not
declared as owl:ObjectProperty in the approved BFO profile
(bfo-2020-without-some-all-times). The well-known part_of (BFO_0000050) and
has_part (BFO_0000051) relations are explicitly excluded from this check because
they are permitted.
Why it matters: The MSE-KG adopts BFO 2020 in a temporally-neutral profile that deliberately omits time-indexed predicates. Using a forbidden temporal predicate would introduce inconsistencies with the chosen ontological commitment and could break reasoning.
Expected result
0 rows. Every row returned identifies a BFO predicate that is not sanctioned by the adopted BFO profile.
##### Forbidden temporal predicate must not be used (e.g. BFO_0000176)
PREFIX owl: <http://www.w3.org/2002/07/owl#>
SELECT DISTINCT ?p WHERE {
?s ?p ?o .
FILTER(STRSTARTS(STR(?p), "http://purl.obolibrary.org/obo/BFO_"))
# skip part_of and has_part, see here: https://oborel.github.io/obo-relations/ro-and-bfo/
FILTER(?p NOT IN (
<http://purl.obolibrary.org/obo/BFO_0000050>,
<http://purl.obolibrary.org/obo/BFO_0000051>
))
FILTER NOT EXISTS {
SERVICE <http://matwerk-virtuoso:8890/sparql> {
GRAPH <https://nfdi.fiz-karlsruhe.de/matwerk/bfo-2020-without-some-all-times> {
?p a owl:ObjectProperty .
}
}
}
}
3 -- Classes Must Be Declared in the Ontology
What it checks:
Identifies any individual that is typed with a class (rdf:type) which is not
declared as owl:Class or rdfs:Class in the reference ontology graph
(mwo/3.0.1), unless that class has been explicitly marked as deprecated
(owl:deprecated true or owl:DeprecatedClass). Built-in OWL classes are excluded
from the check.
Why it matters: Using an undeclared or removed class means the data no longer conforms to the governing ontology. This can happen after ontology upgrades where classes are renamed or retired. Catching these mismatches early prevents silent data loss and reasoning errors.
Expected result
0 rows. Every row returned identifies an instance typed with a class that does not exist (or is deprecated) in the ontology.
PREFIX rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>
PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#>
PREFIX owl: <http://www.w3.org/2002/07/owl#>
SELECT DISTINCT ?s ?c WHERE {
?s rdf:type ?c .
FILTER(isIRI(?c))
FILTER(!STRSTARTS(STR(?c), STR(owl:)))
VALUES ?ctype { owl:Class rdfs:Class }
?c a ?ctype .
# Class must be declared in the ontology graph as owl:Class or rdfs:Class
# Class must not be deprecated in the ontology graph
FILTER NOT EXISTS {
SERVICE <http://matwerk-virtuoso:8890/sparql> {
GRAPH <https://nfdi.fiz-karlsruhe.de/matwerk/mwo/3.0.1> {
{ ?c a ?ctype }
UNION
{ ?c owl:deprecated true }
UNION
{ ?c a owl:DeprecatedClass }
}
}
}
}
4 -- Predicates Must Be Declared Properties in the Ontology
What it checks:
Ensures every predicate used in the graph that is typed as owl:DatatypeProperty or
owl:ObjectProperty is actually declared with that type in either the main ontology
graph (mwo/3.0.1) or the NFDIcore extension graph (nfdicore-extension/3.0.4).
Why it matters: Undeclared properties indicate that the data uses vocabulary not sanctioned by the ontology. This can result from typos, outdated mappings, or incomplete ontology imports. Detecting these mismatches protects the integrity of downstream queries and reasoning.
Expected result
0 rows. Every row returned identifies a property that is used in the data but not declared in any of the reference ontology graphs.
##### Predicates must be declared properties in the ontology
#
# Ensures every predicate we use is declared as a property in the ontology graph.
PREFIX rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>
PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#>
PREFIX owl: <http://www.w3.org/2002/07/owl#>
PREFIX xsd: <http://www.w3.org/2001/XMLSchema#>
SELECT DISTINCT ?p ?ptype WHERE {
?s ?p ?o .
VALUES ?ptype { owl:DatatypeProperty owl:ObjectProperty }
?p a ?ptype .
FILTER NOT EXISTS {
SERVICE <http://matwerk-virtuoso:8890/sparql> {
GRAPH <https://nfdi.fiz-karlsruhe.de/matwerk/mwo/3.0.1> {
?p a ?ptype .
}
}
}
FILTER NOT EXISTS {
SERVICE <http://matwerk-virtuoso:8890/sparql> {
GRAPH <https://nfdi.fiz-karlsruhe.de/matwerk/nfdicore-extension/3.0.4> {
?p a ?ptype .
}
}
}
}
5 -- Unique Textual Entity Values
What it checks:
Detects pairs of distinct resources that share the exact same value for the
nfdi:NFDI_0001007 (textual entity) property. The query returns one row per
duplicate pair, ordered to avoid reporting the same pair twice.
Why it matters: Textual entities in the MSE-KG are expected to be unique identifiers or labels that distinguish one resource from another. Duplicate values suggest either a data-entry error (the same label assigned to two different things) or a merging problem (two IRIs that should have been reconciled into one). Enforcing uniqueness keeps the graph clean and prevents ambiguous lookups.
Expected result
0 rows. Every row returned identifies a pair of resources that incorrectly share the same textual entity value.
# no two "textual entities" may share the same value (NFDI_0001007)
PREFIX rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>
PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#>
PREFIX owl: <http://www.w3.org/2002/07/owl#>
PREFIX xsd: <http://www.w3.org/2001/XMLSchema#>
PREFIX nfdi: <https://nfdi.fiz-karlsruhe.de/ontology/>
SELECT $this ?value ?other
WHERE {
$this nfdi:NFDI_0001007 ?value .
?other nfdi:NFDI_0001007 ?value .
FILTER ( $this != ?other && str($this) < str(?other) )
}
6 -- Disjoint Class Violation
What it checks:
Finds any individual ?x that is simultaneously an instance of two classes ?C and
?D where those classes have been declared owl:disjointWith each other. This is a
classic OWL consistency check that would normally be caught by a DL reasoner.
Why it matters: Disjoint-class axioms encode fundamental domain constraints -- for example, a process cannot simultaneously be a material entity. Violating these axioms makes the graph logically inconsistent, which causes reasoners to derive arbitrary (and therefore useless) conclusions. Running this check without a full reasoner provides a lightweight yet effective consistency safeguard.
Expected result
0 rows. Every row returned identifies an individual that belongs to two mutually exclusive classes, indicating a logical inconsistency.