Let me start by saying I'm very new to the scRNAseq world and am learning my way around R. I now have an scRNAseq dataset and I have identified a subcluster of a cell-type that I'm interested in. I am able to find single markers for this subcluster, using Seurat or sc2marker but I think the single marker approach likely doesn't capture as many cells as could be done through multiple markers. I'd like an approach that maximises sensitivity and specificity for this subcluster, I'm thinking by using a combination of 2+ markers. This would allow me to maximise identification of the number of cells in this subcluster and minimise false-positives (identification of unwanted cell types). At the moment I'm doing this sort of manually and using Nebulosa (joint = TRUE) to view essentially the intersect of my best guess of two markers.
There must be a more systematic way where 1000s of marker combinations are trialled and scored. Is there a tool out there (for R) that can do this? I guess like an in-silico FACS, testing across many combinations.
Hope that was clear.
Thanks!
- Badges
- Science topic