DESeq: Design model for paired design with two treatments
1
0
Entering edit mode
L_K • 0
@l_k-14850
Last seen 23 months ago

Hallo everyone

As I am not exactly an expert in linear modelling and a comprehensive online search did not yield results I'd like to ask for help here:

My experimental design looks as follows: 1 Condition (only individuals with a disease, no controls). Paired treatment design with pre and post treatment, but not the same individuals for both treatments:

My design table thus looks like this:

  Patient1  treatment1  baseline
Patient1  treatment1  treated
Patient2  treatment1  baseline
Patient2  treatment1  treated
...
Patient3  treatment2  baseline
Patient3  treatment2  treated
Patient4  treatment2  baseline
Patient4  treatment2  treated
...


With a total of 6 patients for one treatment and 7 for the other.

My approach now as to first slice the data and seperate the treatments and look for significantly differently expressed genes seperately withing each treatment (pre vs post treatment). This I just did with the Patient+Timepoint (where Timepoint is the 3rd column, pre/post treatment). However, how would I put the design model if I wanted to look for differentially expressed genes between the two treatments taking into account the pre/post treatment status.

Thank you very much for your help!

deseq2 • 419 views
0
Entering edit mode
@mikelove
Last seen 23 hours ago
United States

There is an example in the vignette for just this design, see the section about “group specific condition effects...”

0
Entering edit mode

Thanks a lot. And sorry, my bad!

0
Entering edit mode

Thanks a lot. And sorry, my bad!