Hi all,

I am working on an RNAseq experiment where we have three experimental doses and a control. I have completed the analysis using the dose as a group assignment, but I am curious about genes which have a linear relationship across the gradient of doses. My initial approach was to identify DEGs which changed in the same direction in the pairwise dose vs. control contrasts, but it appears that no genes are represented in all doses. While it might be true that the transcriptional response to each dose is independent, I am wondering if the effect on a gene due to the low dose (relative to the control) might not large enough to be significant in a pairwise comparison, but might be detected as having a linear relationship across doses represented as a continuous variable.

Is there a way to identify genes whose expression has a significant linear relationship with a continuous dose?

Thank you for your help!

If you use the dose as a numeric variable rather than factor, by default you will fit a linear relationship between gene expression and dose.