I was wondering if I understood the
DESeq2 time-series vignette correctly.
My data set contains 5 time-points without control samples as such
> metadata sample time TP1.1 TP1.1 1 TP1.2 TP1.2 1 TP2.1 TP2.1 2 TP2.2 TP2.2 2 TP3.1 TP3.1 3 TP3.2 TP3.2 3 TP4.1 TP4.1 4 TP4.2 TP4.2 4 TP5.1 TP5.1 5 TP5.2 TP5.2 5
So I basically have only the samples in duplicates for each time point and the time as a factor. I would like to identify genes with significant changes of expression over multiple time points.
For that I use the
LRT test and the following model
dds <- DESeqDataSetFromMatrix(countData = counts, colData = metadata, design= ~sample + time) dds <- DESeq(dds, test="LRT", reduced=~time)
Would this give me genes diff. regulated genes over all time-points?
If I want to compare all time-points in a pairwise manner - I should use the
Wald test and the
name parameters in the