W3C home > Mailing lists > Public > public-lod@w3.org > February 2012

Summary of OBO Ontologies Thread (my potentially flawed view)

From: Peter DeVries <pete.devries@gmail.com>
Date: Fri, 24 Feb 2012 18:58:43 -0500
Message-ID: <CAE0MQeEZgsczYHjru+jy9qyVX=Vzt7mK99gS-zO4iPrB5GATXQ@mail.gmail.com>
To: public-lod@w3.org
Cc: Jim Balhoff <balhoff@nescent.org>, Matt <diapriid@gmail.com>, Alan Ruttenberg <alanruttenberg@gmail.com>, Chris Mungall <cjmungall@lbl.gov>, Juan Sequeda <juanfederico@gmail.com>, Kingsley Idehen <kidehen@openlinksw.com>, andy_deans@ncsu.edu, Nathan Wilson <nwilson@eol.org>, David Patterson <dpatterson@mbl.edu>, Anne Thessen <athessen@mbl.edu>
To spare others from having to to through the entire OBO ontology thread, I
thought it would be helpful to make a more digestible summary.

I went to the page that the Hymenoptera Ontology Paper said was the site
for the ontology.

A Gross Anatomy Ontology for Hymenoptera
http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0015991

*"The HAO OBO file is available at
http://www.obofoundry.org/cgi-bin/detail​.cgi?id=hymenoptera_anatomy. "*

That OWL ontology had URI's like this

http://purl.org/obo/owl/HAO_0000144

OBO now has moved to a different scheme at
http://obi-ontology.org/page/Main_Page

Which has URI's like this http://purl.obolibrary.org/obo/HAO_0000144

I imported both the old ontology and the new ontology into my SPARQL
endpoint by placing the URI's for each into my list of imported schema.

At the time it was not clear to me that all of the obo ontologies are
supposed to be at these new URI's and that the Hymenoptera ontology page
above had not been fixed to point to the new ontology.

You can look at one of the current HAO ontologie URI's within the my KB
(virtuoso) http://bit.ly/wNsnBY

( Some details you can skip and come back to later)
* This is after running Kingsley's isparql script to add the
wdrs:describedby links back to the ontology itself.
  This ties each term back to the ontology that defines it. Note in the LOD
it is possible to have a URI and not automatically know where it is defined.
  I also do this when I make a statement about a URI defined somewhere else
so I can get back to the RDF where that statement was made.
  For instance some of the separate RDF or RDFa files that make statement
about GNI names
  e.g Statements about the GNI Name Plasmodium malariae Feletti and Grassi
1889 http://bit.ly/wWN8rW
()

Also note in the KB view there are still blank nodes like *nodeID://b591845
*

This is the result of a links to non-global URIs.

If you look at the ontology file itself you will see that it has axioms on
these terms, but the axioms themselves don't have URI's.

So far we have two issues:

1) The ontology should have wdrs:describedby or rdfs:isDefinedBy?
statements in the owl file.
2) The axioms should have global URI's (I think - Alan might have a
different opinion)

The next issue is, does this vocabulary work on the LOD as it is expected.
(At a minimum following TBL's 4 rules)

e.g. If I ask for the machine interpretable definition of a term - do I get
back machine interpretable information?

One way to test this is to do the following:

curl -H "Accept: application/rdf+xml"
http://purl.obolibrary.org/obo/HAO_0000144

and see if you get back appropriate RDF.

Currently the HAO does not do this. (They are working on it)

Another way to check if it is checking this is to use the Vapour Validator
Service (http://validator.linkeddata.org/vapour)
to see if HAO returns info by any of the proper ways.

If you put in the HAO URI *http://purl.obolibrary.org/obo/HAO_0000144*

It fails,

My comparison if you put in the txn Honey Bee URI *
http://lod.taxonconcept.org/ses/z9oqP#Species*

it works (I thought it would help to see what it looks like when it works)

The Sindice Inspector Tools is also a good thing to use to check this
http://sindice.com/


Issue #3 is that HAO at present does not follow Tim Burners-Lee's 3 rule (
http://www.w3.org/DesignIssues/LinkedData.html)

3) When someone looks up a URI, provide useful information, using the
standards (RDF*, SPARQL)

It looks as if the HAO people are working on fixing this.

In summary.

A) The whole issue of *http://purl.org/obo/owl/HAO_0000144* vs *
http://purl.obolibrary.org/obo/HAO_0000144* is because we were talking
about different files.

B) The things at *http://purl.obolibrary.org/ *are OWL (I will let someone
else say what kind of OWL)

C) There are three issues still outstanding

1) Statements in the OWL file that tie the individual items back to the
ontology (wdrs:describedby or rdfs:isDefinedBy?)
2) The ontologies axioms are blank nodes
3) Properly following TML's 3rd rule.

Why is the 3rd part important? Because when a service sees one of these
terms used it needs to look it up to see how to properly deal with it.

This is the semantic web right?

For instance, Sindice and Sig.ma etc. know that a txn:Taxonomist is a
dbpedia:Scientist which is a foaf:Person

* Where possible I try to define my classes and predicates as subclasses or
subproperties of other well known URI's

I might have written something wrong or potentially debatable above , but
my hope is that this does not start another 50 message thread :-)

Respectfully,

- Pete

------------------------------------------------------------------------------------
Pete DeVries
Department of Entomology
University of Wisconsin - Madison
445 Russell Laboratories
1630 Linden Drive
Madison, WI 53706
Email: pdevries@wisc.edu
TaxonConcept <http://www.taxonconcept.org/>  &
GeoSpecies<http://about.geospecies.org/> Knowledge
Bases
A Semantic Web, Linked Open Data <http://linkeddata.org/>  Project
--------------------------------------------------------------------------------------
Received on Friday, 24 February 2012 23:59:12 UTC

This archive was generated by hypermail 2.3.1 : Sunday, 31 March 2013 14:24:37 UTC