We are hopefully going to release a new version soon with several
enhanced analysis features. In doing so, we have begun to support the
new ontology file format OBO 1.2.

In working with this new format, I have been trying to decide whether
to consider terms that denote regulation of a GO-term (e.g.,
regulation of apoptosis versus apoptosis) as children of that term or
as separate, and thus not a part of the GO category they are
regulating. Currently, they are considered children. If do not
consider these children, this will have a profound effect on which
terms are ultimately reported by GO-Elite.

For example, the GO term calcium-independent cell-cell adhesion
(GO:0016338), has three regulation terms; negative regulation of
calcium-independent cell-cell adhesion, positive regulation of
calcium-independent cell-cell adhesion and regulation of
calcium-independent cell-cell adhesion
(http://amigo.geneontology.org/cgi-bin/amigo/term_details?term=GO:0016338).
In the current version of GO-Elite (version 1.20), these three
"regulation" terms are considered children, however, It is unclear to
me whether this should be the case. If we consider them children, then
the associated genes will be considered a part of (nested with) the
parent (calcium-independent cell-cell adhesion) for pathway
over-representation analysis, otherwise, they will be evaluated
independently. Right now I am leaning towards not including these as
children. My rationale for this is that if one process is regulating
another (e.g., regulation of apoptosis and apoptosis), then these
should be considered independent terms.

Thanks in advance for you feedback.
Best,
Nathan

I agree with your suggestion.

A different consideration/question: If all gene names in 'regulation' are summarized in X, the gene names in 'positive regulation' called Y and in 'negative regulation' called Z. Would X always be equal to Y + Z ? i.e. shouldn't Y and Z then be children of X (instead of 3 seperate processes right now)

  -- Stan

________________________________________
From: go-elite@googlegroups.com [go-elite@googlegroups.com] On Behalf Of Alexander Pico [ap...@gladstone.ucsf.edu]
Sent: 01 June 2011 18:09
To: Nathan Salomonis
Cc: go-elite@googlegroups.com
Subject: GO-Elite User Group Post Re: Question to GO-Elite users

Keeping them as children makes sense. Here's a question that might help clarify the decision. Are most of the "regulators" of a process also members of the process term? If so, then nesting them makes sense. If not, then I would agree with you that these are indeed separate processes and should be not be included as children.

 - Alex

On May 31, 2011, at 9:50 PM, Nathan Salomonis <nsalomo...@gmail.com> wrote: