genes repeated in the heatmap several times(10x) when using limma
2
0
Entering edit mode
@andreilazar9-18021
Last seen 2.9 years ago

Hello,

In the output of the heatmap, there are some repeated genes for several times, and I can't figure out why.  Obvious, something is wrong.

I try to compare a control group (which is in triplicate: C1, C2, C3) with a treated group which is also a triplicate A (A1, A2, A3), and individually, with another treatment group B (B1, B2, B3).  For the sake of example, I had simplified the group's names to letters.

The code that I use for contrast is: 

"comparisonmodel <- model.matrix(~0+design)
> colnames (comparisonmodel) <- levels(design)
 

fit <- lmFit(NmData, comparisonmodel)

#Applying the empirical Bayes method
project.fit.eBayes <- eBayes(fit)
names(project.fit.eBayes)

#Make individual contrasts and fit the new model
CaseControl <- makeContrasts(A-C, B-C, levels=comparisonmodel)
CaseControl.fitmodel <- contrasts.fit(fit, CaseControl)
CaseControl.fitmodel.eBayes <- eBayes(CaseControl.fitmodel)

> #Filtering Results
> nrow(topTable(CaseControl.fitmodel.eBayes, coef=1:2, number=1000, lfc=2))
[1] 925
> probeset.list <- topTable(CaseControl.fitmodel.eBayes, coef=1:2, number=100, adjust.method="BH", sort.by="F", lfc=2)"

Is it possible that the repetition of gens in the heatmap to be because of the contrasts?

Here is a link where you can observe the heatmap. 

https://we.tl/t-4o7TXZw7mB

Thanks

 

 

limma normalization heatmap design and contrast matrix cancer • 503 views
ADD COMMENT
1
Entering edit mode
Aaron Lun ★ 27k
@alun
Last seen 2 hours ago
The city by the bay

You don't actually show the code that you used to make the heatmap, so it's hard to tell what's happening here. I can't see how the effect you're describing would be introduced by limma, I'd suggest you check if you have duplicated entries in NmData itself.

P.S. Don't use lfc in topTable, see the Note in ?topTable.

ADD COMMENT
0
Entering edit mode
Axel Klenk ▴ 980
@axel-klenk-3224
Last seen 10 hours ago
Switzerland

Wild guess (as you don't say where your data come from): some Agilent microarrays contain 10 copies of a subset of their oligos.

ADD COMMENT

Login before adding your answer.

Traffic: 161 users visited in the last hour
Help About
FAQ
Access RSS
API
Stats

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6