I'm a newbie at single cell RNA-seq, and I am using Seurat in R for the analysis. I read that the AddModuleScore function can be used to see the scoring of a set of genes from bulk RNA-seq (signature) in every cell. Then, I did a mean of this signature score of every cell from each cluster.
However, as I understood, these score values should not be compared between the different signatures because they do not have a "physical meaning". Do you know any other way to make the signatures representation comparable between clusters? Someone told me to do a simple average expression, but I am not convinced of statistics behind it, since I have clusters with more quiescent and cycling cells, and these last ones have more counts, so the representation of signatures would be biased.
If you could help me, you would do me a great favour!
This statement says that you cannot compare between two signatures, but you can still compare between clusters. For example the cell cycle signature can be compared across cells and clusters but you cannot compare cell cycle signature with lets says neurogeneic signature.
The other thing you could so is GSVA analysis on pseudobulks of cells from specific clusters.