[BioC] about multtest question #2

[Phguardiol at aol.com]

Hi,
thanks to Rafael and Deepayan I am a little closer to the end but I still=
=20
dont get the identification of the probes differentially expressed using th=
e=20
following with multtest and the doc from the package (I am just following t=
he=20
whole procedure form the documentation of the package using my data from 2=
=20
duplicated experiments I want to compare, there might be optimization for t=
he=20
kind of test to use etc but this is not my purpose right now). Here is the=
=20
procedure and the problem at the end:

library(multtest)
library(affy)

data<-ReadAffy()
data2<-rma(data)
save(data2, file=3D"cwcneo0A.txt")=A0=A0 -> shall I rename it in a differen=
t way=20
using a .rda instead of the .txt ?
class<- c(0,0,1,1)=20

teststat<-mt.teststat(exprs(data2),class)
postscript("mtQQ.eps")
qqnorm(teststat)
qqline(teststat)
dev.off()
pdf("mtQQ.pdf")
qqnorm(teststat)
qqline(teststat)
dev.off()
tmp<- mt.teststat.num.denum(exprs(data2),class,test=3D"t")
num<- tmp$teststat.num
denum<- tmp$teststat.denum
postscript("mtNumDen.eps")
plot(sqrt(denum), num)
dev.off()
pdf("mtNumDen.pdf")
plot(sqrt(denum), num)
dev.off()

rawp0 <- 2 * (1-pnorm(abs(teststat)))
procs<- c("Bonferroni", "Holm", "Hochberg", "SidakSS", "SidakSD", "BH", "BY=
")
res <- mt.rawp2adjp(rawp0, procs)
adjp <- res$adjp[order(res$index), ]
round(adjp[1:10, ], 2)

resT <- mt.maxT(exprs(data2), class, B=3D1000)
ord <- order(resT$index)
rawp <- resT$rawp[ord]
maxT <- resT$adjp[ord]
teststat <- resT$teststat[ord]

mt.reject(cbind(rawp, maxT), seq(0, 1, 0.1))$r
which <- mt.reject(cbind(rawp, maxT), 0.01)$which[ ,2]

Here I am and if I do:=20
geneNames(data2) I get all the probe sets, same with rownames(exprs(data2))
but I just like to have the one identify and "collected" in which !

Thanks for any help
Philippe=20


	[[alternate HTML version deleted]]