Growth and development

[Midori Harris]

Hmm,

Biology has a long, illustrious history of exceptions and of what were 
thought to be unwavering facts being turned on their heads by pesky 
organisms. I would be very hard pressed to come up with a reliable way to 
partition the cell ontology as described in Chris' thought experiment.

midori

On Mon, 25 Apr 2005, Chris Mungall wrote:

> 
> Thanks Jen - I think you've summed it up well.
> 
> Just to reiterate - the ontology should not be constructed around the
> limitations of any particular tool. However, tools can be useful to
> pinpoint areas of the ontology where there is a lack of formal rules and
> consistency.
> 
> Does it not seem slightly unsatisfactory to anyone else that one has to
> appeal to some vague intuition that some particular research community may
> hold to decide where 'x growth' should have an 'x development' parent? It
> seems that this kind of ad-hoc knowledge is fairly fragile and liable to
> change at any time, causing flux in the ontology.
> 
> Or try thinking about it another way: given adequate resources, would it
> be possible to partition the cell type ontology into cells which sometimes
> grow without being part of some developmental program (in non-pathological
> wildtype scenarios) and those that necessarily entail development when
> they grow? Does anyone have a sense of whether that partition could be
> made at a relatively high/generic level in CL, or whether it would be more
> like a collection of ad-hoc exceptions to the
> automatically-entails-development rule? How would that partition differ if
> constructed by a different curator? How stable would this partition be?
> Are there experiments that can be done to justify particular choices in
> constructing the partition?
> 
> This partition - if constructed - could be used by either a computer
> program, or a curator to decide whether x-growth is a child of
> x-development.
> 
> Even if we do not partition CL in this way, it useful to imagine this as a
> thought experiment. Is such a partition even meaningful?
> 
> Here is another, and in my opinion simpler, solution:
> 
> Let's imagine we split growth into 'growth', 'developmental growth' and
> 'non-developmental growth'. Curators can use their judgement to choose
> which cross-products to manifest in the ontology (so there would be many
> more 'x developmental growth' terms than 'x non-developmental growth'
> terms, I would imagine).
> 
> This shifts the question of 'is this instance of growth part-of/is-a
> development?' to the annotator, where it can be decided on a case by case
> basis.
> 
> This can be either good or bad, depending on whether the imaginary
> partition discussed above is a constant, unwavering fact of biology or
> really something that is only true or false on a per experimental
> observation basis.
> 
> Cheers
> Chris
> 
> On Mon, 25 Apr 2005, J Clark wrote:
> 
> > Hi,
> >
> > At the consortium meeting we decided to move 'growth' so
> > that it is a sibling of 'development' rather than a child.
> >
> > Chris and I were talking about it afterwards and he was
> > asking if there is a rule that he can use for obol so that
> > he knows when 'x growth' terms should have an 'x
> > development' parent.
> >
> > For example, is the growth that precedes bacterial division
> > always, sometimes, or never considered to be part of
> > development?
> > I asked around about this, and the view was that some
> > species research communities think that this kind of growth
> > is part of development and some think it isn't. Chris was
> > concerned that if species research communities differed in
> > their view of this then it would be impossible to represent
> > the information in an ontology structure.
> >
> > I am meant to be implementing the changes to the growth
> > terms but I don't feel I can go ahead with that while Chris
> > has these doubts about the representation of different views
> > of growth in a single DAG.
> >
> > After failing to reach a concensus on this at the meeting I
> > thought it would be best to try to resolve this problem as
> > soon as possible while the discussion is still fresh in our
> > minds. I have written to the key people in the discussion to
> > make sure they're all free this week (David, Rex, Tanya,
> > Chris). This e-mail is an attempt to restart that discussion
> > so that Chris can represent his views directly to the people
> > involved. I'm hoping that he can get a satisfactory
> > resolution to his question so I can go ahead and implement
> > the change to the growth term.
> >
> > Thanks for taking the time to help sort this out. I have
> > attached the minutes of the growth v. development discussion
> > in case anybody needs a reminder of what was said.
> >
> > Best wishes,
> >
> > Jennifer
> >
> >
> >
> >
>