Understanding Rdfs And Rdf

Question

[aghili.babak]

If I understand correctly, RDFS is the schema for the RDF. So if it is the schema it means we have no data in it, it just shows the relations, hierarchy,...so for example if I open a RDFS in Protege for example I will be able to navigate to Human Genome-> Human Locus -> Human Gene BUT How about the real data? the genes like ABL1,etc.. so when I am navigating to Human Gene I would like to be able to see all human genes under it. But currently with the RDFS that I have I can't see the actual genes.

And I am confused about that, so data should be loaded from somewhere else? I have everything in excel CSV files. So if I want to see the genes I should create a parser program that reads fmy csv files and injects them inside the RDFS?

I can attach my RDFS for Human genome if that helps .

Thanks.

Answer 1

[Joachim]

RDFS can help you describe relationships and types of RDF instances. Alternatively, there is also OWL which does more or less the same job, but the theory and driving intentions behind both languages are slightly different. A short tutorial about RDFS/OWL can be found here: http://www.linkeddatatools.com/introducing-rdfs-owl

The actual data is encoded as RDF. Have a look at the RDF Primer that also comes with accessible examples such as:

http://www.example.org/index.html has a creator whose value is John Smith

For a bioinformatics example, you can look at the BioPAX documentation, where they are also sharing some recent posters. I think they are good for getting an overview of the topic. Note that BioPAX makes use of OWL, rather than RDFS though.

Live RDF example can be retrieved from resources such as Bio2RDF or the SADI. The latter has an interactive query interface that you can try to get more RDF examples. Go to http://biordf.net/cardioSHARE/ and try this query (then click on "View results as RDF"):

PREFIX sadi: <http://sadiframework.org/ontologies/properties.owl#>
PREFIX ss: <http://semanticscience.org/resource/> 
PREFIX uniprot: <http://lsrn.org/UniProt:> 
SELECT ?nameString
WHERE {
    uniprot:P15923 sadi:hasName ?name .
    ?name ss:SIO_000300 ?nameString .
}

Or, if you are working more on the chemistry side, then you can also get more impressions on the use of RDF in the Chemistry Development Kit in this blog post.

Coming back to your RDFS (and hence, OWL) question: you encode your data in RDF, where RDFS and OWL provide further information on top of your data. Sometimes you find data also encoded in OWL, which is for example the case with the Gene Ontology, but there are a couple of reasons why that happened (they come from another ontology description language originally). If you want to represent data, then focus on RDF and just define what-is-what by linking out your RDF resources to existing RDFS/OWL ontologies. The National Centre for Biomedical Ontology offers a good search interface in their "BioPortal".