Hello Michael and community,

I am using DESeq2 to analyze a time course experiment. My time course experiment has only one (rather than two) group of samples, and I am interested in detecting genes that moved up or down over time in the same way among all samples.

The latest version of DESeq2 tutorial has shown a good example on how to test interaction term between samples and time points. However, there is no illustration on the case of identifying genes with same expression trends (linear or non-linear) over time among multiple samples.

Could anyone give a hint how to do it with DESeq2 or suggest any other robust test?

Thanks!

It sounds like you could transform the data to variance stabilize (e.g. vst), remove the mean per gene across samples, and then cluster the genes.

I don't have any code for this in the vignette or workflow, but it's pretty simple using base R to do this.