W3C home > Mailing lists > Public > public-semweb-lifesci@w3.org > June 2011

Re: [linkedlifew3cnote] Re: Linkset Provenance

From: Helena Deus <helenadeus@gmail.com>
Date: Thu, 16 Jun 2011 15:27:24 +0100
Message-ID: <BANLkTi=3vQU7_9xLoWiWtv6YTEsy0MhUvA@mail.gmail.com>
To: linkedlifedatapracticesnote@googlegroups.com
Cc: Claus Stie Kallesøe <CSH@lundbeck.com>, HCLS <public-semweb-lifesci@w3.org>, Anja Jentzsch <anja@anjeve.de>
Hi Claus, all

Lately I have been preferring the second option, for provenance/attribution
issues  - I want to be able to separate my list of instances of a class from
the list of instances defined somewhere else. A neat way to do that is to
say that
< "drugX" rdf:type myshema:drug > and then somewhere else state that
< myschema:drug owl:sameAs someother:drug >.

A triple store with inference like owlim is able to return "drugX" either
when I ask for ?drug rdf:tpe myschema:drug OR ?drug rdf:type someother:drug

Best
Lena



On Thu, Jun 16, 2011 at 3:18 PM, M. Scott Marshall <mscottmarshall@gmail.com
> wrote:

> Hi Claus,
> [bouncing this to the list because I think this is of general interest and
> there is plenty of expertise there]
>
> This is where many get stuck. Happily, you are not the first one looking
> for specific classes/terms that match your data although there are plenty of
> areas of ontological usage that still need clarification and improvement.
>
> In TMO, we selected from a number of existing ontologies as much as we
> could and filled in where we couldn't find the right fit. See the attached
> image for an overview of the structure. TMO is also available from
> Bioportal.
>
> See also, the NCBO Ontology Recommender as a place to get started.
>
> http://www.bioontology.org/ontology-recommender
>
> http://www.bioontology.org/wiki/index.php/Ontology_Recommender_Client_Examples
>
> -Scott
>
> --
> M. Scott Marshall, W3C HCLS IG co-chair, http://www.w3.org/blog/hcls
> http://staff.science.uva.nl/~marshall
>
> On Thu, Jun 16, 2011 at 3:35 PM, Claus Stie Kallesøe <CSH@lundbeck.com>wrote:
>
>> Ok, now I am ready.
>>
>> I will try to follow your steps so starting by creating an ontology for
>> LudnebckResearch (compound, batch, assay, project, scientist etc).
>>
>> I will use TopBraid composer. Its comprehensive (protege seems even
>> harder) but I think I have worked out what parts I need and from the
>> neologism site I found this - which is actually really helpful in order to
>> understand what it is and what it is not!:
>>
>> ----
>> Neologism is no ontology editor. If you want to build large and complex
>> domain ontologies, if you need the full power of OWL, or if you need
>> sophisticated reasoning services, consider applications such as TopBraid
>> Composer, Protégé, or the NeOn Toolkit. Neologism is designed for the
>> simpler space of RDF vocabularies. OWL ontologies tend to be about modelling
>> a domain with the formal language of logics; RDF vocabularies thend to be
>> about exchanging and integrating data between systems and on the Web.
>>
>> ----
>>
>> So when building an ontology i am thinking of something like
>> lundbeck:compound; lundbeck:scientist etc. As an overall research ontology.
>> I will then link these to the specifics in the relational databases like
>> lundbeck:compound SKOS:narrowmatch/OWL:sameAs ddis:compound (ddis being one
>> of the databases I have mapped with D2R). This is all the Phil approach.
>>
>> But as many of the concepts are known outside lundbeck (and we migth want
>> to link to that later) should I instead use something like this:
>> http://bioportal.bioontology.org/ontologies/45788 as "chemical entities
>> of biologicalinterest" is what I have? Or should I create my own and then
>> later do lundbeck:compound OWL:sameAS chebi:compound or similar?
>>
>> Remember the best practices paper several times mentioned always to
>> reuse....
>>
>> Ok, my last question for today.
>>
>> claus
>>
>>
>>
>> -----Original Message-----
>> From: Anja Jentzsch [mailto:anja@anjeve.de]
>> Sent: Thursday, June 16, 2011 12:51 PM
>> To: Claus Stie Kallesøe
>> Cc: M. Scott Marshall
>>  Subject: Re: Linkset Provenance
>>
>> > I had a look here:
>> > http://www4.wiwiss.fu-berlin.de/latc/toollibrary/example.php
>> >
>> > That might actually be some of the things I have been looking for.
>> > What tools to use for what, in what order and how to put it together.
>> > Anja, this also describes what to do with the link descriptions from
>> > Silk ;-)
>>
>> I wrote this tutorial and it is exactly what I described yesterday ;)
>> Probably I didn't put it specific enough.
>>
>> Anja
>>
>> > I will try to set something like this up to play
>> >
>> > c
>> >
>> > -----Original Message-----
>> > From: Anja Jentzsch [mailto:anja@anjeve.de]
>> > Sent: Wednesday, June 15, 2011 6:02 PM
>> > To: Claus Stie Kallesøe
>> > Cc: M. Scott Marshall
>> > Subject: Linkset Provenance
>> >
>> > I just put this on IRC but you're already gone:
>> >
>> > 17:00 Anja    I wonder if it makes sense to have provenance data on the
>> linksets provided via D2R as a plugin as well
>> > 17:00 Anja    this would cover the concerns Scott was talking about
>>
>>
>
>
> --
> M. Scott Marshall, W3C HCLS IG co-chair, http://www.w3.org/blog/hcls
> http://staff.science.uva.nl/~marshall
>
>


-- 
Helena F. Deus
Post-Doctoral Researcher at DERI/NUIG
http://lenadeus.info/
Received on Thursday, 16 June 2011 14:28:22 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 14:52:47 UTC