Hello,

This is the first time I use Limma for multiple (three) categories comparison. From the tutorial, I use make contrast to compare every two categories. Let's say A,B,C.

contrast.matrix <- makeContrasts("A-B", "A-C", "B-C", levels=designMA)

topTable(rawReadsFit2,adjust="fdr", n=Inf,coef="A-C")

My question is, if I have negative result of logFC for a gene, I know that the gene is a down regulated gene, but which is lower? Is it lower in C or is it lower in A? In other words, If I use A-C, what is the null hypothesis for the gene expression level?

I'm sorry to ask this basic question because if I check the result, it seems a bit reversed from my expectation. I expect that the gene in A should be lower than in C but from the negative result, it means C is lower than A. Probably I used reverse matrix or maybe the result itself is like that. Thank you for your answer.

Your expection is correct. A negative logFC means lower in A than in C. In other words, A - C < 0.