I'm using the timecourse package in bioconductor. I have been obtaining size factors for each of the samples from DESeq's sizeFactors() before log transforming the raw data and using this in mb.long() from timecourse.
Should I be doing this or is timecourse a package which handles the sample size normalization by itself? I ask because I recently re-ran some of my samples forgetting to use the sizeFactor() adjustment and the top genes are much better at reflecting trends seen in other analyses. Which makes me curious.