Re: Advancing translational research with the Semantic Web

Dear All,

Regarding “reification design patterns” and the reification & OWL (not 
the thorny logic-based representation of beliefs et al), permit me to 
mention that support for n-ary relations ---where n may also be >2--- in 
description logics is already possible with DLR [1] and implemented with 
reasoner-support in the iCOM tool (the tool may not live up to 
end-user-level expectations on userfriendliness, but it works) [2]. In 
theory, this could also be added to OWL without increasing complexity 
and would at least relieve modelers from the burden of doing 
reifications themselves.


Staying with article comments but going off topic from this thread, 
there were two aspects in particular that read oddly and raises more 
questions than that it provides information. One about the BioRDF, the 
other the ontology working group.

First, on BioRDF: ``BioRDF has the goal of converting a number of 
publicly available life sciences data sources into RDF and OWL.'' This 
sounds very much like preparing for use of Semantic Web technologies as 
opposed to using them. Given the many legacy data and the need for 
mappings, is there an intention to do something about standardization, 
structuring, or coordination of those transformations? (e.g., there are 
several efforts on OBO-to-OWL mappings).

Second, the ontology working group: ``A goal of the HCLSIG is to 
facilitate creation, evaluation and maintenance of core vocabularies and 
ontologies to support cross-community data integration and collaborative 
efforts. Although there has been substantial effort in recent years to 
tackle these problems, the methodology, tools, and strategies are not 
widely known to biomedical researchers.'' I’ve read the majority of 
emails on this list, but have not found a clear set of methodologies, 
tools, and strategies from which an interested biomedical researcher can 
compose her own cocktail (and where did the ontology maintenance aspect 
go that was in the task group list of activities last year?). Further, 
``The role of the ontologies task force is to work on well-defined use 
cases, supporting the other HCLSIG working groups.'', which does not 
sound as if one is going to reach those unaware biomedical researchers 
to convert & train them (although Matthias Samwald’s linking up with the 
SW outreach he mentioned yesterday seems a promising step in that 
direction.). Also, if I were a biomedical researcher, I would not be 
convinced by 1-3 use cases, but maybe they are (maybe set up a 
literature database on experience and implementations?); I also doubt if 
they would be sufficient for drawing a set of guidelines.

Last, mere curiosity: ACCP wants temporal construct and constraints, 
which are not properly supported in RDF, OWL and reasoners, but that 
does not reappear in the “identified technical limitations” section. 
Given that adding rules has its own issues as well, this leaves logic 
programming (from the list mentioned in the ACCP section). But do you 
want to have proper support of the temporal stuff in a logic-based 
ontology language and reasoner in some not-too-distant future, or be 
fine with other means? The article suggest the latter.

[1] this is an overview article: Calvanese, D., De Giacomo, G. 
Expressive description logics. In: The Description Logic Handbook: 
Theory, Implementation and Applications, Baader, F., Calvanese, D., 
McGuinness, D., Nardi, D., Patel-Schneider, P. (Eds). Cambridge 
University Press, 2003. pp178-218.
[2] latest version downloadable from 
http://www.inf.unibz.it/~fillottrani/ontoeditor.zip and some earlier 
background information: http://www.inf.unibz.it/~franconi/icom.

Best regards,
Marijke

C. Maria Keet
KRDB Research Centre
Faculty of Computer Science
Free University of Bozen-Bolzano
Piazza Domenicani 3
39100 Bozen-Bolzano
Italy
tel: +39 04710 16128
fax: +39 04710 16009
email: keet@inf.unibz.it <mailto:keet@inf.unibz.it>
web: http://www.inf.unibz.it/krdb/
home: http://www.meteck.org <http://www.meteck.org/>

samwald@gmx.at wrote:

>  
>
>>><protein_a_expression_process> <has_participant> <protein_a> .
>>><protein_a_expression_process> <located_in> <tissue_b> .
>>><protein_a_expression_process> <described_by> <source_c> .
>>>      
>>>
>>The "protein expression process" class that needs to be introduced here 
>>does seem a bit like bending over backwards 
>>    
>>
>[...]
>  
>
>>also you need to introduce
>>an 
>>additional predicate, "has_participant". This would certainly work,
>>however 
>>>from my point of view the best solution is the solution that uses the 
>>fewest custom concepts, 
>>    
>>
>
>Hi Eric,
>
>The choice of this design pattern is not arbitrary, it is based on the OBO Relation Ontology [1], BFO and the OWL version of the Gene Ontoloy, which are becoming widely accepted. Other foundational ontologies (like DOLCE) have similar relations and entities. 
>In other words: I have not introduced this design pattern to conform to my suggested approach. A well-designed ontology would use such classes and relations right from the start. Using such a design pattern increases opportunities for interoperability between ontologies significantly.
>
>cheers,
>Matthias Samwald
>
>----------
>
>Yale Center for Medical Informatics, New Haven /
>Section on Medical Expert and Knowledge-Based Systems, Vienna /
>http://neuroscientific.net
>
>  
>

Received on Friday, 18 May 2007 08:16:40 UTC