[BioC] Basic question re: multiple testing

[Adaikalavan Ramasamy]

If you use Bonferonni correction, the ranking does not change (strictly
speaking when adjusted p-value < 1), because p_adj = min( 1, n*p ). Ie a
multiplicative constant n is applied. This is a very conservative
estimate as it assumes all the genes are independent of each other. 

You can learn a lot more about the actual formulae by looking into the
codes of p.adjust(). 

pp <- runif(10000)
ppa <- p.adjust( pp, method="fdr") 
plot(pp, ppa)


-----Original Message-----
From: Claire Wilson [mailto:ClaireWilson at picr.man.ac.uk] 
Sent: Friday, May 16, 2003 9:35 PM
To: BioC mailing list
Subject: [BioC] Basic question re: multiple testing


Dear all,

Apologies if this is slightly off-topic/bit basic! With respects to
multiple testing, if you apply it does it change the actual order of
significance, i.e. are your top 10 most significantly changing genes the
same when you perform a t-test with multiple testing and when you apply
a t-test without multiple testing?

Many thanks in advance

Claire

--
Claire Wilson   
Bioinformatics group  
Paterson Institute for Cancer Research  
Christies Hospital NHS Trust  
Wilmslow Road,  
Withington  
Manchester  
M20 4BX  
tel: +44 (0)161 446 8218  
url: http://bioinf.picr.man.ac.uk/
 
--------------------------------------------------------

 
This email is confidential and intended solely for the use of th... {{dropped}}