I'm using DEXSeq to analyze some RNA-seq data. Despite reading the manual I'm not quite sure I'm doing it correctly for my paired samples. I have for samples, with the current design for the ExonCountSet: condition experiment Sample1 treatment 1 Sample2 treatment 2 Sample3 control 1 Sample4 control 2 So the first two samples are the experimental ones, the two last ones controls, paired with the first two. In the ensuing analysis, I do as follows: formula1 = ~ sample + exon + experiment:exon + condition:exon formula0 = ~ sample + exon + experiment:exon ecs = estimateDispersions(ecs, minCount=5, nCores=cn, formula=formula1) ecs = testForDEU(ecs, nCores=cn, formula0=formula0, formula1=formula1) Is this the correct way to account for the paired nature of my samples? [[alternative HTML version deleted]]

Hi Philip On 10/01/14 15:45, Philip Jonsson wrote: [...] > formula1 = ~ sample + exon + experiment:exon + condition:exon > formula0 = ~ sample + exon + experiment:exon > > ecs = estimateDispersions(ecs, minCount=5, nCores=cn, formula=formula1) > ecs = testForDEU(ecs, nCores=cn, formula0=formula0, formula1=formula1) > > Is this the correct way to account for the paired nature of my samples? Yes, this is correct. One point: Make sure that 'experiment' is interpreted as a factor and not as an integer vector. Simon