I recently performed RNA-seq of 24 samples. 12 samples were healthy and 12 samples had been exposed to a chemical. For each of the 24 samples, the titer of this chemical was measured 24 hours after administration (and just before the RNA was extracted).
I got a rather small list of DEGs (say n=100) between the two treatment groups. When I examine the 24 sample read counts for each DEG, I notice that certain samples really drive the differential expression (about 4 samples have very large values). I also notice that these 4 samples had high titer values.
My question is: How can I design a statistical test to determine if any of these DEGs correlate statistically to the recorded titers?
The (exaggerated here) data would look something like this:
Titer values = [1, 2, 1, 1, 2, 1, 2, 3, 2, 1, 2, 1, 3, 2, 1, 11, 12, 1, 13, 14, 1, 2, 2, 1] (four samples are high)
DEG1 = [3, 5, 6, 1, 2, 4, 3, 5, 4, 4, 3, 4, 3, 2, 5, 34, 25, 1, 37, 53, 1, 3, 2, 4] (same four samples are high) .... DEG100 = [...]
Any ideas would be greatly appreciated. If you believe I should post this on another board, please let me know. So far, I am only posting it here. Thank you!
