I have data from a Tn-Seq experiment. We have a barcoded knockout library. We measured the abundance of each knockout at time zero (by taking an aliquot from the common library), then did a 2x2 factorial experiment with 5 replicates per level.
I understand how to use (the excellent) DESeq2 to analyze the factorial structure of my data, by having a two-factor colData data.frame and using a formula with an explicit interaction (e.g. ~ A + B + A:B). This is very useful for telling me how the treatments differ in their effect on relative abundance.
However, it does not incorporate the T0 data. If I could incorporate that, it would also give me information on absolute fitness, which would be very useful for my study.
I have an alternate in-house pipeline that first calculates fitness by finding log2 changes from T0 for each replicate, then just does a two-factor lm on each gene's fitness with subsequent BH corrections, but I think this in-house pipeline doesn't do a great job normalizing gene counts and is giving spurious results.
Is there a way I can do this in DESeq2? I apologize for any lack of clarity. Thanks!