[BioC] about understanding contrast matrix

[yanju at liacs.nl]

Dear Bioconductors,

In analyse of DNA microarray using LIMMA: for example you have a contrast
matrix like follow:
---makeContrasts(B-A,C-B,levels=c("A","B","C"))
then it is fitted to linear model and eBayes function is used to summarize
the statistics.
Finally log odds is used to rank the differential genes.

I am wondering is the log odds rank the genes which express different
between condition B and A (B-A), or C and B(C-B), or different in both
conditions, or in either conditions? how to explain it?

Expecting your reply!

Regards,
Yanju