A Brief on ICBO 2009

Some 200 participants enjoyed 3 days of biomedical ontology at the first ICBO conference, preceded by 4 days of classes and tutorials on topics ranging from spatial ontology and qualitative reasoning to the metaphysical foundations of biomedical ethics. The meeting itself included 38 presentations, 43 posters, 6 software demonstrations, 2 panels, and 1 plenary lecture (by Howard Garner, on the use of high-powered text-mining software applied to PubMed for drug discovery).

The meeting covered applications of ontology ranging from chemistry and molecular biology to clinical medicine and from microarray experiments to multi-species anatomy, with particularly noteworthy presentations from Mike Bada, Chris Mungall, and Melanie Courtot, whose talk introduces a new verb ('to mireot') into the language. The papers are available both on the ICBO site as a single document, and also as a special collection on the Nature Precedings site, which includes posters and powerpoint presentations.

ICBO consisted entirely of plenary sessions, and one particularly noteworthy feature of the meeting was the degree to which the very same people who were enjoying talks on the classification of lipids in the morning were sitting, rapt, in the very same room enjoying talks on genome-wide association studies or on physiological simulation in the afternoon. Ontology, along exactly the lines which some of us have predicted for many years, is hereby providing the glue to meld together not just data and information but also specialist communities of human beings, including communities otherwise not known for cross-disciplinary collaboration.

The glue melds not because of special features of ontology software or of ontology formal languages, or because the different collections of concepts in the minds of individuals schooled in different specialisms become somehow magically compatible when these individuals are placed in a room where ontology is being discussed. Rather, the glue melds because ontology provides a new, common way to look at reality -- of lipids, of gene-disease associations, of physiology -- the very same reality which the different scientific specialisms are all, in their different ways, concerned to examine.

Perhaps the most interesting talk at the meeting was that of Harold Solbrig from the Mayo Clinic on the topic of "Concepts, Confusion and Modeling". As Solbrig, points out:

'The term "concept" continues to be used in models to reference categories, classes, universals, individuals and other less well defined artifacts' in a way which 'obfuscates the purpose and usefulness of the model itself', leading some to propose the total banishment of the term from ontological discourse.

As Solbrig reports in documenting his own impressions of the meeting:
'One of the emerging issues in the ISO world centers on the IT definition of "ontology" as 'a description from someone's perspective of what is'. This causes folks to confuse the description with what is being described and leads them to make bizarre assertions like "there are many ontologies, therefore there are many realities". I equate it to taking the word "chemistry" and using it to describe chemistry books. Suddenly, we've all got our favorite "chemistry" -- and universities can begin to collect and catalog chemistries from a variety of sources.  Unfortunately, we have now lost a word to describe what all these "chemistries" are about, and may even get involved in nonsensical arguments like "My chemistry only has 103 elements, while yours has 118.'

On the realist approach, which Solbrig now defends, ontology is not about books, or theories, or descriptions, or concepts, or data, or information. Rather it about the entities themselves, out there in reality. It is about the very same entities which scientists are investigating empirically in the lab and in the clinic. And it works, as many contributions to the ICBO meeting are beginning to demonstrate, precisely to the degree that this realist orientation is given full play.

-- Barry Smith, National Center for Biomedical Ontology, National Center for Ontological Research, University at Buffalo