[BioC] Help with raw p-value

Sandrine Dudoit sandrine@stat.berkeley.edu
Sat, 19 Oct 2002 20:46:03 -0700 (PDT)


Hi,

The p-values in rawp0 are calculated based on a standard normal
distribution, while those in t.test are calculated based on a
t-distribution with df = 11.049.

Try
2*(1-pt(abs(teststat[1]),df=11.049))
and you will get the same p-value as with t.test.

Sandrine

-------------------------------------------------------------------------------
Sandrine Dudoit, Ph.D.               E-mail: sandrine@stat.berkeley.edu
Assistant Professor                  Tel: (510) 643-1108
Division of Biostatistics            Fax: (510) 643-5163
School of Public Health		     http://www.stat.berkeley.edu/~sandrine
University of California, Berkeley
140 Earl Warren Hall, #7360
Berkeley, CA 94720-7360
-------------------------------------------------------------------------------

On Wed, 16 Oct 2002 donghu@itsa.ucsf.edu wrote:

> I have played with multtest package using Golub data.  After loading the
> dataset, I followed the steps descripted in the document to calculate t
> stat and raw p-values.  Here are what I did:
>
> >teststat <- mt.teststat(golub, golub.cl)
> >rawp0 <- 2*(1-pnorm(abs(teststat)))
>
> Then I took a look of the t stat and raw p-value of the first gene.  I got:
>
> >teststat[1]
> [1] 1.759195
> >rawp0[1]
> [1] 0.07854436
>
> I also tried to calculate t stat and p-value of the first gene using
> t.test.  I did the followings.
>
> > x <- golub[1,1:27]
> > y <- golub[1,28:38]
> > t.test(x,y)
>
> The result showed "t = -1.7592" and "p-value = 0.1062".  I don't
> understand why the "p-value" I got here is different from "rawp0[1]"
> although the t stat is same.  Can someone let me know why?  Thank you in
> advance.
>
> Donglei Hu
> Gene Array Core Lab
> Diabetes Center
> UCSF
>
>
> _______________________________________________
> Bioconductor mailing list
> Bioconductor@stat.math.ethz.ch
> http://www.stat.math.ethz.ch/mailman/listinfo/bioconductor
>