Confusion about Limma group design

Question

[zizigolu]

Hi

I have done differential expression of microarray data by below code

> fit <- lmFit(data, design)
> contr <- makeContrasts(Responder - Non_Responder, levels = colnames(coef(fit)))
> contr
               Contrasts
Levels          Responder - Non_Responder
  Responder                             1
  Non_Responder                        -1
> tmp <- contrasts.fit(fit, contr)
> tmp <- eBayes(tmp)
> top.table <- topTable(tmp, sort.by = "P", n = Inf)
> head(top.table, 20)

Now, for a given gene, I got confused if this gene has been upregulated in Responder group or Non_Responder group

For instance logFC for APC5 in totable is 0.215476927, APC5 has been up regulated in Responder group or Non_Responder group

Answer 1

[ATpoint]

The contrasts, e.g (Responder - Non_Responder) always read as Responder over (that is divided by) Non_Responder. The first group is the counter, the second one the denominator. Therefore positive logFC indicates overexpression in Responder. Still, be sure not only to look at logFC but also at the FDR since FC alone is prone to bias and/or large variation between replicates and therefore requires the statistics to be considered reliable (or not).