Hi,
I would like to clearly understand some points of edgeR in order to use it with flow cytometry counts. I focussed on pre-filtering (filterByExpr()) and extracting frequencies (cpm()).
a) I would like to set the same threshold/prior from one experiment to another and express them in counts per million. This view is maybe stupid, let me know.
At the moment, my aim doesn't seem direct. Correct me if I'm wrong:
b) filterByExpr is using a min.count but min.count is relative to the median library size, not to 1e6 ==> edit: correct, clearly described in help(filterByExpr)
c) cpm() is using a prior but the prior is relative to the mean library size, not to 1e6
d) filterByExpr is using cpm() but no prior can be passed and the default (2) is used ==> edit: cpm() is used without log, so prior is ignored
e) edgeR is currently be used for analyzing cytometry counts. The diffcyt package applies glmLRT. Should I use glmQLFTest instead?
Thanks for reading.
