W3C home > Mailing lists > Public > public-semweb-lifesci@w3.org > August 2007

RE: HCLS Demo at ISMB/ECCB, How to contribute to the demo?

From: Nigam Shah <nigam@stanford.edu>
Date: Tue, 14 Aug 2007 12:34:09 -0700
To: "'Michel_Dumontier'" <Michel_Dumontier@carleton.ca>, "'Matthias Samwald'" <samwald@gmx.at>, <brandizi@ebi.ac.uk>, "'Eric Neumann'" <eneumann@teranode.com>
Cc: "'W3C HCLSIG hcls'" <public-semweb-lifesci@w3.org>, "'M. Scott Marshall'" <marshall@science.uva.nl>
Message-ID: <000801c7deaa$165abb40$431031c0$@edu>

>Fantastic material Nigam (and co)!  I hope you don't mind if I use
some
>for my 4th year bioinformatics class :-)

By all means...

>Couple of points:
>
>Problem: Asserting disjointness between classes that are definable.
>Example (slide 100): Rural Area and Urban Area are disjoint; Woomera
is
>initially asserted to be a rural area, and it is later develops into
an
>urban area. How do you intend to manage this change? Will you then
>create a new instance of Woomera, assert it as an urban area, and
link
>it to the rural Woomera? Better to not make these disjoint....

We don't handle time in this tutorial at all...(mainly to keep life
simple). I don't have an easy fix for the case that you raise. The
example is intended to explain the meaning of disjointness in an
intuitive manner.

>Problem: Ontology normalization.
>Slide 108: "An InfectiousPneumonia is either a BacterialPneumonia or
a
>ViralPneumonia." An infectious pneumonia is a pneumonia caused by an
>infectious agent. Rather that asserting the union of specific agent
>caused pneumonia, one should instead define a class of infectious
>agents. Thus, we can automatically discover whether the agent is in
fact
>infectious. Moreover, this is easier to maintain.

Again, the Pneumonia ontology is only for teaching the basics of OWL
in under 2 hrs. In fact, if you read the definitions closely, there
are a few mistakes built in (which the classifier will later find
;-)). The assumption that InfectiousPneumonia is either a
BacterialPneumonia or a ViralPneumonia is mainly to keep the tutorial
ontology small. If it did it for a hospital, what you say is
absolutely correct.

>Slide 115: providing definitions for property characteristics would
be
>useful here - the difference between functional/inverse function may
not
>be clear from these examples (as a SIN is also functional - at least
in
>Canada)

Point taken  (didn't make much sense is EU either ;-)).

Regards,
Nigam.
Received on Tuesday, 14 August 2007 19:34:59 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 18:00:49 GMT