POC observations so far

[Toby Kellogg]

Hi all -
  Felipe has made great progress here downloading the various programs
necessary for viewing and manipulating the ontologies.  He's run into some
mechanical problems that he can sort out with Pankaj, the most critical
being trying to get two DagEdit windows open at the same time so two
ontologies can be viewed side by side.  In terms of the ontologies
themselves, we immediately found a number of terms that are not in common
use, or are in odd hierarchical relationships; these should be easy enough
to change.
  A more interesting and complex issue comes with species-specific terms.
An easy one is "silique" in Arabidopsis, which would be an instance of
"fruit" if one is working with multiple species.  A more difficult one is
"stigma", which is a part of "pistil", part of "gynoecium", part of
"floret", etc. up to tassel.  Unfortunately, stigmas do not form in
tassels, because the gynoecium stops developing.  Similarly, "abscission
zone" is part of "silique" in Arabidopsis, and would end up being part of
"fruit" if "silique" were interpreted as an instance of "fruit".  However,
abscission zones do not form in the grass fruit so couldn't be a part of
fruit.  in both cases we end up violating the True Path Rule.  It may be
that this is inevitable, since the descriptors aren't strictly
hierarchical.   Obviously one can get around this somewhat by creating
species-specific bits of the hierarchy, and by making creative use of
"sensu"; this will probably work fine as long as the ontology only has to
apply to Brassicaceae and Gramineae.  If the long-term goal is to make it
apply to all flowering plants, though, there may be a limit to how
species-specific we make the ontologies.   For example we could divide
fruits into indehiscent and dehiscent and then have abscission zone as part
of dehiscent fruits, which would be OK until we get to a fruit that forms
an abscission zone but doesn't dehisce.  Another possibility that Felipe
and I explored a little would be to add another category of connection,
such as "a process that can occur in"  - in addition to "is part of", "is
an instance of" and "develops from".  I suspect that another category might
create more problems than it solves, but it seemed worth considering.
  Any thoughts on this are welcome!
Toby

Elizabeth A. Kellogg
Department of Biology
University of Missouri-St. Louis
8001 Natural Bridge Road
St. Louis, MO 63121
phone: 314-516-6217
fax: 314-516-6233
http://www.umsl.edu/divisions/artscience/biology/Kellogg/Kellogg/