The GO ontology is set up as a directed acyclic graph, where a parent
term is comprised of all its child terms. If you do a standard
hypergeometric, you might e.g., find 'positive regulation of kinase
activity' to be significant.
If you then test 'positive regulation of catalytic activity', which is
parent term, then it might be significant as well, but only because of
the terms coming from positive regulation of kinase activity.
The conditional hypergeometric takes this into account, and only uses
those terms that were not already significant when testing a higher
order (parent) term.
For a reference, see the paper by Adrian Alexa. You can find the
citation on the last page of the 'How to use GOstats' vignette.
Chanchal Kumar wrote:
> Dear Bioconductor developers and users,
> I am using the "GOstats" package to find over/under enriched GO
> in my dataset. And there is an option to calculate the "conditional
> hypergeometric test". I am not sure what this would imply, I am
> the conventional hypergeometric test but this is a bit unfamiliar
> concept. Therefore it will be very helpful if someone could explain
> concept and point me to relevant references.
> Thanks in advance!
> Best Regards,
> Chanchal Kumar, Ph.D. Candidate
> Dept. of Proteomics and Signal Transduction
> Max Planck Institute of Biochemistry
> Am Klopferspitz 18
> 82152 D-Martinsried (near Munich)
> e-mail: chanchal at biochem.mpg.de
> Phone: (Office) +49 (0) 89 8578 2296
> Fax:(Office) +49 (0) 89 8578 2219
> Bioconductor mailing list
> Bioconductor at stat.math.ethz.ch
> Search the archives:
James W. MacDonald, M.S.
Affymetrix and cDNA Microarray Core
University of Michigan Cancer Center
1500 E. Medical Center Drive
Ann Arbor MI 48109