Re: [BioRDF] BioSIOC / aTag task

Matthias and Kei,

This is an interesting idea which Matthias already brought up in the  
Scientific Discourse group - but I am not sure how it is meant to  
align with SWAN-SIOC, if at all.

Ideally we should not needlessly create another ontology of scientific  
discourse in Bio-RDF distinct from the one we are working on in  
Scientific Discourse.  No problem if there are deficiencies in SWAN- 
SIOC which we need to remedy or unaddressed use cases, but let's  
ensure we are building on what is already there.  Perhaps I don't  
understand Matthias' proposal or its use case, but somehow it strikes  
me as potentially falling in the realm of what we have already built,  
and I would like to probe this issue further.

Matthias, do you see your "high-level view of biomedical knowledge in  
the form of small statements"  as something uniquely different from  
SWAN "research statements"?    Would the statements have discourse  
relationships amongst themselves?

Is there a particular reason you feel it would not work to re-use or  
extend SWAN-SIOC  to achieve what you have in mind? If so, how would  
what you propose relate to and align with the work already done in  
SWAN-SIOC?

Best

Tim






On Jan 6, 2009, at 9:14 PM, Kei Cheung wrote:

>
> Hi Matthias,
>
> What you proposed sounds like an interesting project to be included  
> as part of the BioRDF task force. I like the idea of exploring a fit  
> between folksonomy and ontology. Currently, the BioRDF group has the  
> query federation project that explores how to pose federated queries  
> across multiple data/knowledge sources including the HCLS KB's  
> (hosted at DERI and Free Univ. Berlin). The possibility that the  
> BioSIOC project can add more data/knowledge to the HCLS KB hosted at  
> DERI is complementary to the query federation project.
>
> I look forward to working with you.
>
> Cheers,
>
> -Kei
>
> Matthias Samwald wrote:
>
>>
>> In the coming weeks / months I will create some RDF/OWL resources  
>> that make
>> use of the basic SIOC vocabulary together with biomedical domain  
>> ontologies
>> for the representation of biomedical knowledge (let's call this  
>> task "BioSIOC", since I intend to do it as part of the BioRDF task  
>> force [1]).
>>
>> The basic idea is to represent a high-level view of biomedical  
>> knowledge in
>> the form of small statements, represented as SIOC items that are  
>> annotated
>> with entities from OBO ontologies and UniProt entries. So instead of
>> representing the full semantics of the biological statements in RDF/ 
>> OWL
>> (with the associated problems of complexity and inter-dataset  
>> heterogeneity
>> that  made it difficult to create practical applications out of the
>> biomedical RDF/OWL resources that have been created in recent  
>> years), the
>> statements will be represented a small snippets of text and sets of
>> associated tags ('aTags'). Finding the right trade-off between  
>> expressivity
>> and simplicity while optimizing for usefulness in realistic  
>> biomedical
>> research scenarios will be a major focus. The project will be  
>> documented on
>> a wiki page [2].
>>
>> Through the use of SIOC, these statements can also be integrated  
>> into simple
>> (micro-) blogging environments or more specialized and sophisticated
>> discourse databases such as those described with the SWAN ontology.  
>> I will
>> also be working on developing user interfaces for creating,  
>> embedding and
>> querying such SIOC items.
>>
>> If the other participants of the BioRDF task agree, the RDF/OWL  
>> that is
>> produced by this project will also be added to the HCLS Knowledge  
>> Base
>> hosted at DERI [3] (which currently solely consits of the  
>> Neurocommons Knowledge
>> Base). The technologies and data developed by this project will  
>> also be
>> employed for the creation of a Drupal-based knowledge repository  
>> that will
>> be hosted by the Konrad Lorenz Institute for Evolution and Cognition
>> Research [4, 5] (which I will describe in more detail at a later  
>> time).
>>
>> [1] http://esw.w3.org/topic/HCLSIG_BioRDF_Subgroup
>> [2] http://esw.w3.org/topic/HCLSIG_BioRDF_Subgroup/aTags
>> [3] http://hcls.deri.org/
>> [4]
>> http://en.wikipedia.org/wiki/Konrad_Lorenz_Institute_for_Evolution_and_Cognition_Research
>> [5] http://kli.ac.at/
>>
>> Cheers,
>> Matthias Samwald
>>
>> DERI Galway, Ireland
>> http://deri.ie/
>>
>> Konrad Lorenz Institute for Evolution & Cognition Research, Austria
>> http://kli.ac.at/
>>
>
>
>

Tim Clark
Director of Informatics, MassGeneral Institute for Neurodegenerative  
Disease
Instructor in Neurology, Harvard Medical School
617-947-7098 (mobile)

Received on Wednesday, 7 January 2009 03:03:27 UTC