[BioC] Help with limma experiment design - two color data

[Yair Benita]

Hi All,
I have two color microarray data (genepix format) from patients with cancer.
I have a variable number of replicates per patient and in some cases a
dye-swap. On each slide we compared the RNA from tumor tissue to RNA from
healthy tissue in the same patient.

Going through the limma documentation I used the approach below but I am not
sure its correct and makes sense. I am looking for genes differentially
expressed between healthy and tumor tissues. The approach below resulted in
12,000 significant genes after multiple testing correction. I am wondering
if anyone could suggest an approach that makes most sense.

Finally, can someone tell me how to export data for making a heatmap? With
Affy its always easy to just export the expression values of significant
genes, but what do I export here to see expression level of each significant
gene on each slide?

Thanks for the help,
Yair

#dataset was background corrected (rma), normalized within arrays and
normalized between arrays (quantile)

#setup design matrix with replicate information
pateint1<-c(1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0)
pateint2<-c(0, 0, 1, 1, -1, -1, 0, 0, 0, 0, 0, 0, 0, 0)
pateint3<-c(0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0)
pateint4<-c(0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0)
pateint5<-c(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, -1, -1)

design<-cbind(pateint1, patient2, patient3, patient4, patient5)

#setup contrast matrix to find difference between control and tumor
cont.matrix<-makeContrasts(CONTROLvsTUMOR=(pateint1+patient2+patient3+patien
t4+patient5)/5, levels=design)

#fit linear model
fit<-lmFit(MA, design)
fit2 <- contrasts.fit(fit, cont.matrix)
fit2<-eBayes(fit2)
topTable(fit2, adjust="BH")