Entering edit mode
Hi, I have a group of genes upregulated in various clusters of cells when I run Seurat's FindAllMarkers() function. Is it statistically correct to use run compareCluster::compareCluster() on a list of these upregulated genes grouped per cell cluster? Or must the genes be grouped per GO before running compareCluster().
Thanks in advance for your kind help.
Not sure whether I am answering your question, but the function
compareCluster()
will automagically check whether the input genes have a GO annotation (through the annotation info inorg.Hs.eg.db
(when working with human genes)). Those genes that don't have a GO annotation will be discarded from the analysis. Then a (simple) over-representation analysis will be performed for each cluster (=set of selected genes).