[BioC] Multtest

Stephen Henderson s.henderson at ucl.ac.uk
Wed Mar 22 19:25:31 CET 2006


I don't much use multtest itself preferring the built in methods that go
with limma (e.g. toptable). I think however you would be better using
the 'mt.rawp2adjp' function.
e.g
## using Benjamini-Hochberg (or something like that)

> mt.rawp2adjp(p, "BH")->rej
> rej$adjp[1:3,]

             rawp           BH
[1,] 3.245434e-17 1.477904e-12
[2,] 5.406144e-17 1.477904e-12
[3,] 8.744990e-15 1.593774e-10

>cutoff<-.05

## i.e which in the the BH column do you want
> wh<-which (rej$adjp[,2]< cutoff)

## the adjp list is sorted the index to genes is in rej$index 

> ind <- rej$index[wh]

## you want something like

cbind(geneNames(myEset)[ind], rej$adjp[wh,])-> output

## or use geneNames(myEset)[ind] to query the annotation packages for
symbols, refseqs etc.


Cheers
Stephen Henderson
Wolfson Inst. for Biomedical Research
Cruciform Bldg., Gower Street
University College London
United Kingdom, WC1E 6BT
+44 (0)207 679 6827


-----Original Message-----
From: bioconductor-bounces at stat.math.ethz.ch
[mailto:bioconductor-bounces at stat.math.ethz.ch] On Behalf Of Roth,
Richard
Sent: 21 March 2006 19:59
To: bioconductor at stat.math.ethz.ch
Subject: [BioC] Multtest

Hi Stephen, I have started using the package 'multtest', and I have it
working, but I have a few questions which probably show I don't really
know how the package is working!  1. The following commands will provide
a count of probe sets that fall into each defined category and create a
vector of probe sets that meet a specified criteria, but I don't
understand what the $r or $which are doing on the end of these commands,
I realize they are necessary, but...
>mt.reject(cbind(RawpvT,maxT),seq(0,1,0.1))$r &
which<-mt.reject(cbind(RawpvT,maxT),0.01)$which[,2].  2. Is there an
easy way to get a list of the probe sets and their associated p-values
(raw and adjusted) that meet a specified criteria in a single data
frame/matrix?

Thanks, Rich


Rich Roth, PhD
Senior Scientist, Bioinformatics
Molecular Medicine
Neurocrine Biosciences
858-617-7204



**********************************************************************
This email and any files transmitted with it are confidentia...{{dropped}}



More information about the Bioconductor mailing list