[BioC] finding GO terms that are close in the GO graph

[Karin Lagesen]

I have managed to generate two sets of GO terms, set A are the GO
terms for about 60 genes, and set B are the GO terms for around a 1000
genes. My goal with this is to find genes from the larger set that
have similar GO terms to those mapped to the genes in the smaller
set. I have preliminarily decided to go hunting for genes in the
larger set whose go terms are either the same, a parent, a child or
share the same GO term parent as a gene in the smaller set. (I hope
this makes sense!)

So, I am wondering how to do this. I was thinking of plainly getting a
list of all parent, child and siblings of GO terms in set A, and using
external means of matching that up to those found in set B. Are there
any methods out there that either would do this, or do something else
that would help me achieve this goal?

I have looked at quite a few of the GO packages (GOSemSim, GOstats and
others), but I couldn't figure out if there were methods in those
packages that could be used for this purpose.

TIA,

Karin
-- 
Karin Lagesen, Ph.D.
karin.lagesen at medisin.uio.no
http://folk.uio.no/karinlag