We used the van Elteren test implementation for Seurat objects (https://github.com/KChen-lab/stratified-tests-for-seurat) to detect marker genes for each node of the cluster tree, comparing the current cluster either against all other nodes on the same level or against the sibling nodes to obtain a set of markers per cluster. We subset the tree into its neurons and non-neurons and calculated the global marker genes only versus other cells in the respective subtrees to restrict the marker sets to more relevant genes.
Do you have any questions about this protocol?
Post your question to gather feedback from the community. We will also invite the authors of this article to respond.