Finding DEGs of cell type comprising of more than one clusters

Question

[sc_analysis]

When one cluster represents one cell type we can use FindAllMarkers and find the DEGs of one CLUSTER in comparison to rest of the CLUSTERS and then annotate the cell type to each cluster based on canonical cell type markers and can cell that the DEGs are deferentially expressed in different clusters. But what to do when more than one cluster represents one cell type? How to find the deferentially expressed genes in one CELL TYPE (comprised of more than one cluster) in comparison to rest of the CELL TYPES? For example cluster 1,3,7 represent osteoblasts cells, then how to find the DEGs of osteoblasts in my dataset in comparison to rest of the cell types?

[Haci]

You can just create a new column in your meta data table and aggreate your clusters of interest into one level (leaving the rest), and then perform DE within this new column as you described in your first sentence.

Answer 1

[Alex Gibbs]

In theory, if your 3 clusters represent the same cell type. Then running FindAllMarkers using the 'seurat_clusters' as idents should show enrichment of expected genes for all 3 cluster DEGs. I would reccommend using featureplots to observe expression of known cell type markers in conjunction with the DEG analyses.