W3C home > Mailing lists > Public > public-semweb-lifesci@w3.org > September 2012

Re: Our white paper on Semantic Node Labeling in Clinical models

From: Michel Dumontier <michel.dumontier@gmail.com>
Date: Mon, 17 Sep 2012 18:29:51 -0400
Message-ID: <CALcEXf4V11gJFHapGzW+NouVBDUtS+dqEMS3UzuP0Lpf3=ZhTQ@mail.gmail.com>
To: Peter.Hendler@kp.org
Cc: ratnesh.sahay@deri.org, eric@w3.org, helena.deus@deri.org, kerstin.l.forsberg@gmail.com, LINMD.SIMON@mcrf.mfldclin.edu, meadch@mail.nih.gov, mscottmarshall@gmail.com, public-semweb-lifesci@w3.org
Hi Peter,
   here's my view on it:

1. The goal of a terminology (at least in this context) is to provide terms
to specify types and relations that enable the construction of meaningful
statements.

2. The goal of an information model is to specify and restrict the number
and type of relations along with their allowed_subjects/objects (beyond
what would be specified in some terminology).

I appreciate, as you've stated in prior discussions, that you're just
fighting the first battle which is to separate the "what" from the rest of
the info. Yet, it's entirely feasible to include all aspects of a clinical
event (what, where, who, why, when, how) using terminological resources.
 We've heard about the Salus project, which seems to push everything into
RDF for query, and from which we can then leverage referred to ontologies
for enhanced query answering provided some minimal reasoning capability
(which you can do upfront if desired).

Meanwhile, the role of the information model is really to guide systems
(and their users) to construct and view a potentially limited set of
meaningful statements, whether to meet data entry expectations,
compatibility with production systems, or to conform to standards for data
exchange. SPARQL can be used to check whether minimal conditions are being
met, or marshal data into a common structure, or otherwise serve up errors.

In your semantic node proposal, you propose a number of "tags" for
labelling nodes:


   - vocabulary (any standard vocabulary)
   - intensional logic  (RDF, OWL-DL, EL, EL+, SNOMED CT, etc.)
   - ontology (SNOMED CT, etc.)
   - hierarchies (Clinical Findings, Observables etc.)
   - post-coordination-allowed (TRUE / FALSE)

but i didn't recognize sufficient justification for all of these tags. It
seems to me that you really need to specify i) a list of allowed terms and
ii) whether to include their descendants. Is the "intensional logic" tag
really necessary ? Seems like meta-information about the terminological
choice in the first place. Perhaps you can provide a counter to this
simplified proposal.

m.



-- 
Michel Dumontier
Associate Professor of Bioinformatics, Carleton University
Chair, W3C Semantic Web for Health Care and the Life Sciences Interest Group
http://dumontierlab.com


On Mon, Sep 17, 2012 at 5:19 PM, <Peter.Hendler@kp.org> wrote:

> This very clearly (to me at least) explains the problem and suggested
> solution to some of the problems we (Kaiser) see in clinical models.
>
> Wonder who agrees?
>
>
>
>
>
>
> *NOTICE TO RECIPIENT:*  If you are not the intended recipient of this
> e-mail, you are prohibited from sharing, copying, or otherwise using or
> disclosing its contents.  If you have received this e-mail in error, please
> notify the sender immediately by reply e-mail and permanently delete this
> e-mail and any attachments without reading, forwarding or saving them.
>  Thank you.
>
>
>
Received on Monday, 17 September 2012 22:30:39 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 18:01:14 GMT