UK Ontology Network: from Archipelago to Pangaea

At a meeting of minds at the KMI, Open University on Friday 30th April, 2010, it was decided the UK ontology community should attempt to coordinate its efforts in a slightly more structured (I hesitate to use the word ‘formal’ ) manner. The participants, including both myself and Janna Hastings from EBI, represented a variety of organisations, both inside and outside of academia, from areas including bioinformatics, health care, chemistry, formal ontology, ontology languages, geographical surveying, philosophy, engineering and computer science.

The topography of the UK
Image via Wikipedia

It appeared to me that there was a general consensus in the room that present ontology work in the UK appears as an archipelago; formed of multiple, disparate efforts, often without knowledge of things occurring in organizations up and down the land. The desire is to better coordinate and collaborate, as is the true spirit of ontology work; who coordinates the coordinators indeed. We need to connect these islands and form our pangaea.

The outcome of the meeting was the establishment of a UK Ontology Network, whose members are those working in ontology and ontology related projects based in the UK. As work continues on creating a more permanent web presence (this is one of my tasks so I can say with confidence it will be done ‘soon’), a mailing list has already been established. For those interested, please subscribe to the Google Group to join us in our efforts to give a stronger cross-domain community feel to the UK based ontology work that we all know goes on, though we may not know exactly what and where, or by whom. Hopefully we will know  soon.

JM

Reblog this post [with Zemanta]

The Fast and the Furious?

Dietrich Rebholz-Schuhmann gave a talk at the Ontology Interest Group today on using ontologies within the contact of text mining. It provided a lot of perspective and I think gave a strong and convincing use case as to how ontologies could be used. It also poses questions as to how amenable to such analysis ontologies are. For instance, those developed in upper ontologies which often use abstract terminology such as those seen in BFO do not reflect normal biological language. Relations such as has_quality for relating phenotypic information are unlikely to produce results used on current biomedical literature. There is also a suggestion that an increased number of more granular relationships can help to improve this analysis. Minimally, richer synoyms for both classes and relationships would help both text mining analysis and human understanding.

My favourite quote of the talk:
“Ontology: The Slow and the Furious”.

JM

BBC One-tology

There was a recent posting to the OntoGenesis network I thought it would be worth sharing. It highlights some of the work done at the BBC on wildlife linked data at the new Wildlife finder.

This work was done by Tom Scott and Leigh Dodds et al and can be found at http://purl.org/ontology/wo/

It appears the BBC are using the ontology to generate rather attractive web pages on various taxa, e.g. http://www.bbc.co.uk/nature/class/Amphibian

and by adding ‘.rdf’ to the end of each link, a dereferenceable rdf fragment is visible:
http://www.bbc.co.uk/nature/class/Amphibian.rdf

This is not the first time the BBC have looked into using ontologies, although it is clearly the most relevant to bioontologies. They have also spend time on developing a ‘programme ontology‘ which is also publicly available. The BBC are not alone in large organisations investing time into ontology usage. Last year, NASA announced it too had been using ontologies to perform automated data analysis and for organising earth and environmental terminology. Perhaps the kill aspp is on its way;  Google-ontology…

JM

Next Meeting of the Cambridge Ontology Reading Group

The next meeting of the Ontology Reading Group will take place on the 11th of March at 8pm in the Kingston Arms (Kingston Street).

Please post suggestions for reading material either to the mailing list or as comments to the blog. I’ll circulate a final agenda a week before the meeting, and as mentioned before, we won’t expect much in terms of advance preparation.

We can also try to use the opportunity to informally discuss potentials for expanding our local collaborations?

Cheers, Janna

ICBO 2011

Congo Buffalo
Image by Durotriges via Flickr

Barry Smith announced at the OBO Foundry workshop this week that the 2nd International Conference on Biomedical Ontologies will be held in Buffalo, New York, US on the July 29th-31st 2011. Note, this confirms that there will not be an ICBO 2010.

JM

Reblog this post [with Zemanta]

OWL vs OBO

Red Stag Deer Locking Horns
Image by ifijay via Flickr

Thanks to Duncan Hull and Midori Harris for their talks today on OWL language and OBO format. The merits of both have been discussed, argued and fought over previously on several other online fora so it is probably not worth me recapitulating all of these. Instead I will point to some useful pages online and leave it to others to detail. From my own personal perspective, and without revealing any preferences, I would say that it is my belief the OWL specification driven approach has greatly benefited from the OBO format user driven approach and vice versa. It is also clear to me we are still learning from each other even now.

Phil Lord: OBO Format to Manchester Syntax

Carole Goble and Chris Wroe (2004). The Montagues and the Capulets Comparative and Functional Genomics, 5 (8), 623-632 DOI: 10.1002/cfg.442

Allyson Lister: OBO to OWL and back again

(please add others as comments to this blog)

JM

Reblog this post [with Zemanta]

Announcing the Third Meeting of the Ontology Interest Group @EBI

The next meeting of the EBI ontology interest group will take place on Thursday, 11 Feb at 11:00 am in A2-33. During the meeting we will talk about ontology languages:

  • Duncan Hull (Chemoinformatics and Metabolism Team) will give an overview over the Web Ontology Language
  • Midori Harris (GO Team) will do the same for the OBO Language, comparing and contrasting the two.

Looking forward to seeing you all there!

Nico

Announcing the First Meeting of the Ontology Reading Group

The  first Reading Group meeting will take place next thursday at 7pm at the Kingston Arms.

For reading material, let’s all read these two papers, which have similar topics:

Integrating phenotype ontologies across multiple species. Genome Biology
DOI:10.1186/gb-2010-11-1-r2

Linking Human Diseases to Animal Models Using Ontology-Based Phenotype
Annotation. PLoS Biol 7(11): e1000247.
DOI:10.1371/journal.pbio.1000247

*beforehand* please ;-).

The third suggested paper we’ll carry over to our meeting next month.

All the best,

Janna

Second Meeting of the Ontology Interest Group, 28/01/09

Today we had the second meeting of the ontology interest group at the EBI. The programme was as follows:

Robert Hoehndorf: “What is an Ontology?”
Short use case 1: James Malone: “Ontologically Modeling Sample Variables in Gene Expression Data”

Short use case 2: Dietrich Rebholz-Schuhmann: “Extraction of GO terms from literature for the identification of gene-disease associations”

The meeting was very well attended with around 30 people. Robert’s talk gave a great overview over the various uses and usages of both the term and the tool/methodology/technology that “ontology” represents and the evolutuon of “ontology” as a discipline in both philosophy and computer science and artificial intelligence as well as the relationship between them. He has promised to write up his talk into a blog post soon. Until he does so, here are his slides:

“What is an ontology”

James then presented a use case, showing how  ontologies, such as the Experimental Factor Ontology could be used for sophisticated querying of resources such as Array Express Atlas and data integration.

Unfortunately, Dietrich didn’t get to present his talk as we ran out of time during the meeting: it will be the first item on the agenda for the next meeting in approximately 2 weeks.

During the meeting it became apparent, that we should talk about ways of formalizing ontologies, i.e. ontology languages next rather than to start with ontology development best practice. In particular, the group seemed to be interested in OWL. We will therefore aim to start with OWL in the next meeting.

There have also been suggestions from several members of the group to preserve the talks as videos and to make them publicly available, which is a great idea and we will look into this.

So far the rough summary of the meeting. As always, please comment, correct any mistakes, carry on the conversation in the comments section of the blog. I will send a link to a poll to determine the best time for the next meeting via the mailing list.

Nico

Blogging an Ontology Book

University of Manchester
Image via Wikipedia

I and fellow EBIers Duncan Hull and Helen Parkinson have just completed a two-day meeting at Manchester University titled “blogging a book”. The idea was the brain child of Phil Lord and was ran by Robert Stevens and David Shotton under the OntoGenesis Network. The aim of the meeting was to produce encyclopedic-like articles about ontology (particular ontologies in biology) which might be of use to those coming to the field from new (or perhaps just after a refresher or something to cite).

The articles produced include topics such as :

What is an Ontology?
Upper ontologies

OWL

RDF

Application and reference ontologies

Community dirven ontology development
Protege and Protege OWL
Semantic Integration in Life Sciences

As well as producing interesting and hopefully useful articles, a further motivation of the meeting was to see if the authoring and peer-review process could be duplicated in an online, blog-style framework.  Each blog article, following authoring, is submitted for review and will only be considered ‘published’ when it has passed reviews from 2 peers.  It exposed some interesting strengths and weaknesses of authoring technology such as blogs.

Perhaps the primary weakness of the approach was the lack of collaborative software for authoring an article between multiple collaborators, although this is generally not how blog articles are written. Perhaps this is an area that may be strengthened in future. A strength is that the articles were quickly published and that DOI will (eventually) be obtained for articles so that they are citeable and therefore credit can be assigned.

Perhaps a more debatable strength, is that reviews were posted alongside the finished article for all to see, including the reviewers names, rendering the process entirely transparent. It raises some interesting thoughts, not least, would a reviewer be more or less inclined to criticise a paper they were reviewing if their identity was revealed and if their review were to be made public? Would reviews be more thorough as a consequence, since they would, in essence, also be published and available for peer scrutiny?

James Malone