I just wanted an honest opinion on whether to include two groups together or analyse separately in DESeq2 (please see the pca plot using the link below). I am inclined towards analysing separately, but I want to hear from the readers.
Here, the two groups represent two tissues and within each tissue there are three sub-groups (three breeds)
Hi Mike, sorry for incomplete post. The study consist two breeds and their F1 (i.e 3 breed groups) and two tissues. The two groups you can see in the PCA is for two tissues. One tissue (E) is more heteregenous than the other (C). Within E, there is high within-group variability. I have updated the post but I will also upload another PCA.
Is there additionally batch information? It looks like the right group is split 3/3, 3/3/ and 2/4 by something other than what you've got being used to color now.
If you suspect batches, you could use sva or RUVSeq to identify them. See the last section of our workflow.
If not, then yes, I'd recommend splitting into a C dataset and an E dataset to perform DE analysis separately, as they will have very different estimates of variability.
You need to post more information. How can you perform differential analysis with less than two groups?
Hi Mike, sorry for incomplete post. The study consist two breeds and their F1 (i.e 3 breed groups) and two tissues. The two groups you can see in the PCA is for two tissues. One tissue (E) is more heteregenous than the other (C). Within E, there is high within-group variability. I have updated the post but I will also upload another PCA.