How To Best Get Readable Obo Rdf Notation3 Serialization?

Question

[Egon Willighagen]

The OBO ontologies in OWL have the tendency to have cryptic rdfs:Class names, like the Information Artifact Ontology: IAO_0000116. However, this makes RDF serializations hard to read (as Notation3):

ex:CDK rdfs:subClassOf iso:IAO_0000010 .

This may be overcome by adding the label in the output:

ex:CDK rdfs:subClassOf iao:IAO_0000010 .
iao:IAO_000010  rdfs:label "software" .

Another alternative is to make a shadow ontology and define:

custom:Software owl:equivalentClass iao:IAO_0000010 .

And then use in the rest of the serializaton:

ex:CDK rdfs:subClassOf custom:Software .

So, my question here is to what practices are used in current bioinformatics projects to work with OWL versions of the OBO Foundry ontologies? Is the project just hiding all the RDF triples, or are the visible to the user (at some level) and you have the GUI use rdfs:label's instead of URIs? Are there alternative solutions to make serialized RDF using OBO ontologies more human readable?

Answer 1

[Nico Adams]

Egon, the only solution, as far as I can see is the one that Protege takes: use the rdfs:label entries to display human readable class-names rather than the actual class names. By the way - it's not JUST the OBO ontologies in OWL that use this practice: using a numberical class name insulates against changing class names if, down the road, it has been felt that the initial choice of class name was not fortuitous. It's not just the OBO people who do this. The upside is greater term stability, the downside is, that it is damn hard to write software to this. Incidentally and as a tangent - owl:sameAs is only used to assert equivalence between Individuals. If you want to say that classes are the same, you should use owl:equivalentClass

[Egon Willighagen]

Fixed s/sameAs/equivalentClass/