Dear all,

I am new to RNASeq analysis and have found the relevant packages that exist in bioconductor really useful.
I have Readcounts from RNA-seq experiments each having 3 replicates.

         Genotype Treatment Time_hrs Replicate
BM1_WTUT       WT        UT        0         1
BM2_WTUT       WT        UT        0         2
BM3_WTUT       WT        UT        0         3
BM1_WTi8       WT  Stimuli1+Stimuli2        8         1
BM2_WTi8       WT  Stimuli1+Stimuli2        8         2
BM3_WTi8       WT  Stimuli1+Stimuli2        8         3
BM1_WTi2       WT  Stimuli1+Stimuli2        2         1
BM2_WTi2       WT  Stimuli1+Stimuli2        2         2
BM3_WTi2       WT  Stimuli1+Stimuli2        2         3
BM1_WT8        WT       Stimuli1        8         1
BM2_WT8        WT       Stimuli1        8         2
BM3_WT8        WT       Stimuli1        8         3
BM1_WT2        WT       Stimuli1        2         1
BM2_WT2        WT       Stimuli1        2         2
BM3_WT2        WT       Stimuli1        2         3
BM1_FLUT  KO        UT        0         1
BM2_FLUT  KO        UT        0         2
BM3_FLUT  KO        UT        0         3
BM1_FLi8  KO  Stimuli1+Stimuli2        8         1
BM2_FLi8  KO  Stimuli1+Stimuli2        8         2
BM3_FLi8  KO  Stimuli1+Stimuli2        8         3
BM1_FLi2  KO  Stimuli1+Stimuli2        2         1
BM2_FLi2  KO  Stimuli1+Stimuli2        2         2
BM3_FLi2  KO  Stimuli1+Stimuli2        2         3
BM1_FL8   KO       Stimuli1        8         1
BM2_FL8   KO       Stimuli1        8         2
BM3_FL8   KO       Stimuli1        8         3
BM1_FL2   KO       Stimuli1        2         1
BM2_FL2   KO       Stimuli1        2         2
BM3_FL2   KO       Stimuli1        2         3

I would like to test the effect(s) of different Treatments and different time points on genotype. As you see in the table, I have two levels in Genotype, three levels in Treatment and 3 levels in Time_hrs.
In my design, I have used:

dds_genotype.treatment.2h <- DESeqDataSetFromMatrix(countData = ReadCounts.new.genotype.treatment.2h, colData = colData.new.genotype.treatment.2h, design=~Genotype + Treatment + Genotype:Treatment)

and

res <- results(dds, contrast=c("Genotype","KO", "WT"), pAdjustMethod = "BH")

But I guess this does not allow me to distinguish between the potential impact of treatment at different time points on genotype.

Could you please suggest me of potential ways of analyzing these data? The linear model has to incorporate all of these variables I guess.

Many thanks,
Dimitris

This is a sufficiently complex experiment that I think if you are not sure what design to use, which contrasts to build and how to interpret these, I'd recommend you collaborate with a local statistician. There are more than one way to setup the design, and a number of possible contrasts to use representing different null hypotheses and comparisons of interest, and the interpretation of these is best explained in person. So it is more statistical support than software support. Any person with a statistical or linear modeling background can help here: there is nothing unique to DESeq2 about the terms in the model, or the contrasts, it is the same terms and contrasts you would test in a normal linear model.