Differential Expression between 3 groups ?

Question

[hafiz.talhamalik]

How can i perform differential gene expression for 3 groups sample ? edgeR uses t-test which can be used for 2 groups. So how can I do it for 3 groups ? my experiment design is like Normal vs disease condition 1 vs disease condition 2

Answer 1

[ATpoint]

What question do you want to answer? Common DE genes between the two diseases? If so do groupwise comparisons between the diseases and the control and then check for overlaps. You can also go for the average disease effect (disease1 + disease2)/2 - Normal. Check the edgeR manual. There are plenty of examples on how to make contrasts based on the underlying question. See section 3.2.4 on page 32 (Questions and contrasts) of https://www.bioconductor.org/packages/release/bioc/vignettes/edgeR/inst/doc/edgeRUsersGuide.pdf

[hafiz.talhamalik]

yes i want common DE genes. Thanks for the answer.

Answer 2

[ResearchR]

I never used edgeR, but in DESeq2 you can specify multiple conditions as factor in the model formula. What you maybe could do, is to run test like "C1 vs Normal" and "C2 vs Normal" and see which genes appear to be diff. expressed in both conditions. Of course, it might happen that genes are also diff. expressed between C1 and C2, which you have to take in consideration. In addition, multiple test ingis an issue as well. I had a similar question for differentially methylated regions and found an appropriate solution here: DMRSeq

[hafiz.talhamalik]

That could be the one solution to do like this. but we have anova test for multiple sample testing. I m more interested in that if we could apply that for rna-seq analysis..

let me check DMRseq package as well.