[BioC] GOHyperG stats question

[Suraj Menon]

Thanks, Jim - that was quite useful.
Here is a little something for your amusement-
http://papers.ssrn.com/sol3/papers.cfm?abstract_id=946199

Cheers
-Suraj


>>> "James W. MacDonald" <jmacdon at med.umich.edu> 18/07/07 2:30 PM >>>
Hi Suraj,

Suraj Menon wrote:
> Hi all,
> 
> Just a simple question on how GOHyperG calculates hypergeometric p-values - 
> is the test one-tailed or two-tailed? 
> and what would be the reason for using either of them?

The test is one-tailed. See ?HyperGParams-class, in particular the 
testDirection slot.

In general you would want to use a one-tailed test because you are 
testing a particular hypothesis (e.g., certain GO terms are 
over-represented) rather than a more general hypothesis (e.g., certain 
GO terms are represented at a level not consistent with their abundance 
in the universe from which they were selected). In addition, a 
one-tailed test will be more powerful.

I suppose one might want to do a two-tailed test if there really were no 
hypothesis being tested, but instead the goal was to see "what's going 
on" in a certain experimental system.

Best,

Jim


> Cheers,
> 
> 
> Suraj Menon
> PhD Student
> Department of Pathology
> Henry Wellcome Building
> School of Medicine
> Cardiff University
> Heath Park
> Cardiff CF14 4XN
> 
> Tel:    +44 29 2074 3979
> Email: MenonS1 at Cardiff.ac.uk
> 
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at stat.math.ethz.ch
> https://stat.ethz.ch/mailman/listinfo/bioconductor
> Search the archives: http://news.gmane.org/gmane.science.biology.informatics.conductor

-- 
James W. MacDonald
Affymetrix and cDNA Core
University of Michigan Cancer Center
1500 E. Medical Center Drive
7410 CCGC
734-647-5623


**********************************************************
Electronic Mail is not secure, may not be read every day, and should not be used for urgent or sensitive issues.