logFC in limma
1
0
Entering edit mode
gv ▴ 40
@gv-6516
Last seen 17 days ago
United States

I have the following limma code which I wrote(dummy)

The matrix is designed in a way so that I don't use the intercept part.

library(limma)
d <- matrix(rexp(200, rate=.1), ncol=9,nrow = 10)
colnames(d) <- c("A_1","A_2","A_3","B_1","B_2","B_3","C_1","C_2","C_3")
design <- model.matrix(~0 + factor(rep(1:3, each=3)))
colnames(design) <- c("A","B","C")
fit <- lmFit(d,design)
contrast.matrix <- makeContrasts(A-B,B-C,A-C,A-(B+C)/2, levels = design)
fit2 <- contrasts.fit(fit, contrast.matrix)
fit2 <- eBayes(fit2)


In the output of topTable if logFC is positive for contrast A-B does it mean it is log(A/B)?

limma logFC • 365 views
3
Entering edit mode
@james-w-macdonald-5106
Last seen 2 days ago
United States

If it's positive, it means the average for group A is larger than group B (e.g., up-regulated in A vs B), and vice versa. The sign of the statistic doesn't change the relationship. In other words contrast A-B is log(A/B), regardless of the sign.