Re: [HCLS] Bridging Ontologies - with Foundational Ontologies

[William Bug]

Yes - as you and Amit point out, there are many very successful applications and meta-analyses that have been performed with "isolated" ontologies. Where the difficultly comes in is when you want to: a) interlink the results from these separate studies b) formulate additional hypotheses based on these studies What you begin to confront is the fact that if the ontologies were isolated both in terms of not using a shared foundational ontology, there were likely also isolated in the sense of not sharing "best practices" for ontology construction: - many definitions are imprecise, circular, or - in fact - non-existent  - the same is true for relations - they are used in an imprecise, inconsistent, or incommensurate manner (a big problem initially for GO) - mixing "is_a" with the MANY KINDS of "part of" can be a semantic data integration nightmare**. - loosely defined and used lexical facets associated with nodes in the ontology - e.g., acronyms, abbreviations, and misspellings may get used indiscriminately as synonyms - sometimes that's valid, and sometimes its not. - plurals and pre-coordinated descriptions cutting across knowledge domains may appear in the graph - which can limit computability at fine-grained levels. I know this sounds like I'm speaking in the abstract but believe me, I'm not.  There are specific examples where all of these issues have come up for us on the BIRN ontology efforts.  As Peter Mork mentioned for caBIG, registering with a query mediator requires mapping source data models to ontology-base semantic declarations.  The problems above can be a big problem if more than one ontology is required that have overlapping domains.  These inconsistencies need to be disambiguated or reconciled in some way, and in doing so, you may invalidate mappings made to data model elements. This means you no longer have an explicit statement about semantics, but an implied one based on the knowledge in the head of the annotator.  This is also true if you are using algorithmic lexical parsing of some sort, as you are eventually parsing against terms applied by a human in a "loose" manner. Don't get me wrong.  It may sound like I'm slamming GO here.  I'm not.  I've great admiration and respect for their experience and their efforts.  Most commendable of all is the fact that they've gradually learned a lot - both from the school of hard knocks and from the C.S., library science, and ontological philosophy fields that can helped them to develop better shared practices both for ontology creation, curation, and use.  To my mind, the 25 slides prepared by Suzi Lewis & Michael Ashburner (slides 11 - 37) in the following presentation summing up the hard won lessons of the GO Consortium do a better job at establishing such guidelines than anything else I've seen anywhere: Principles for Building Biomedical Ontologies http://www.geneontology.org/teaching_resources/presentations/2005-10_ISMB_Ontology-Building_various.ppt I'd specifically draw attention to the procedural flowchart on slide 19 as being an excellent foundation from which to build more specific guidelines, some of which can be culled from other portions of the presentation.  There have also been many successes that have made it very clear, ontologies are an extremely important part of biomedical KE/KR/KD and semantically-driven data federation. Cheers, Bill **As an example of  On Aug 22, 2006, at 8:51 AM, Kashyap, Vipul wrote:   But - it's not clear to me whether we'll be able to evolve highly automated semantically-formal neuroinformatics analysis systems.  I'm not thinking of reasoning oriented systems, but simply analysis of semantic info a la the ubiquitious use of Gene Ontology in the bio-molecular informatics world.   [VK] It’s interesting to see the ground covered by Gene Ontology without the use of foundational ontologies   Cheers,. Bill   On Aug 22, 2006, at 7:49 AM, Kashyap, Vipul wrote:     Great to hear that! It really seems that most of the promises of semantic web ontologies are only realised when top-level ontologies like DOLCE are used. Maybe we should evaluate the potential use of DOLCE or BFO for the BioRDF tasks?   [VK] Whereas I agree with the use of foundational ontologies, I may not agree with the sweeping generalization above. Significant potential can be realized by using not so formally organized resources such as the UMLS for instance.   ---Vipul     Bill Bug Senior Research Analyst/Ontological Engineer   Laboratory for Bioimaging  & Anatomical Informatics www.neuroterrain.org Department of Neurobiology & Anatomy Drexel University College of Medicine 2900 Queen Lane Philadelphia, PA    19129 215 991 8430 (ph) 610 457 0443 (mobile) 215 843 9367 (fax)     Please Note: I now have a new email - [EMAIL PROTECTED]       This email and any accompanying attachments are confidential. This information is intended solely for the use of the individual to whom it is addressed. Any review, disclosure, copying, distribution, or use of this email communication by others is strictly prohibited. If you are not the intended recipient please notify us immediately by returning this message to the sender and delete all copies. Thank you for your cooperation. Bill Bug Senior Research Analyst/Ontological Engineer Laboratory for Bioimaging  & Anatomical Informatics www.neuroterrain.org Department of Neurobiology & Anatomy Drexel University College of Medicine 2900 Queen Lane Philadelphia, PA    19129 215 991 8430 (ph) 610 457 0443 (mobile) 215 843 9367 (fax) Please Note: I now have a new email - [EMAIL PROTECTED] This email and any accompanying attachments are confidential. This information is intended solely for the use of the individual to whom it is addressed. Any review, disclosure, copying, distribution, or use of this email communication by others is strictly prohibited. If you are not the intended recipient please notify us immediately by returning this message to the sender and delete all copies. Thank you for your cooperation.