Hello,

Suppose I have a model with 4 tissue types A, B, C, D and I want to test the differences of the differences. If I use a cell means model and the following code:

results(dds, contrast = c(1,-1,-1,1))

I was wondering how to interpret the output. In particular, does the log fold value give me log(A)-log(B)-log(C)+log(D) = log((A/B)*(D/C))? Setting lfcThreshold then gives a threshold on the above value?

Thank you!

You can't interpret the output directly, but instead you should plot all the significant genes you care to interpret. An alternative that I have used in the past is to use kmeans clustering on the set of significant genes, and then use spaghetti plots to visualize the expression patterns for the K groups.

And yes, the lfcThreshold is based on the logFC, which is log(A)-log(B)-log(C)+log(D).