Deal with the time-course RNA-Seq datasets

Question

[wangdp123]

Hi there,

I have a set of time-course RNA-seq datasets.

The experiment design is as follows:

1. For each condition, there are 3 replicates.
2. In total, there are 5 time points: T0; T1; T2; T3; T4.

I notice that there is a tutorial from this website (http://master.bioconductor.org/packages/release/workflows/vignettes/rnaseqGene/inst/doc/rnaseqGene.html#time-course-experiments) that explains how to perform time-course analysis for a control and treatment time series.

However, in my case, there is no control-treatment. I wonder how to make the design in DESeq2 right and if there is a need to consider the replicates as a variable in the design?

Many thanks,

Tom

Answer 1

[andrew.j.skelton73]

If all you're interested in (or have to work with), is the differences between time points, then that's all you can test in your experiment. Section 9.6.1 of the Limma users guide is a good basis to start from. These principles can be applied to DESeq2 too. There are other ways to look at your data, including using time as a linear variable, or splines... but for now differential expression between time points is the best starting point.