Dear all, I have an scRNA dataset that I wish to perform differential gene expression analysis on using Seurat. My question relates to the output of FindAllMarkers().
Procedure
I have subset the cells in the Seurat object based on their expression (or not) of an exogenous gene:
> poscells <- WhichCells(object, expression = exogenousgene > 0)
> object$exogenous_exp <- ifelse(colnames(object) %in% poscells, "transfected", "non-transfected")
> exogenous.split <- SplitObject(object, split.by="exogenous_exp")
I make a new assay "integrated" then run SelectIntegrationFeatures(), FindIntegrationAnchors() and IntegrateData(), followed by RunUMAP(), FindNeighbors() and FindClusters().
I then switch back to my "RNA" assay and attempt to look for all marker genes:
> DefaultAssay(combined) <- "RNA"
> genes <- rownames(x = combined)
> combined <- ScaleData(combined, features = genes)
> markers <- FindAllMarkers(combined, grouping.var = "exogenousgene", verbose = TRUE)
Output
My issue is here; when I look at the markers identified, many are listed with numbers appended to the end (e.g., JCHAIN.2 JCHAIN.8 JCHAIN.9, rather than simply JCHAIN). I believe these numbers represent the clusters in which the genes were identified, which is then compared with every other cluster to compute p_val and avg_log2FC.
> markers
JCHAIN.2
JCHAIN.8
JCHAIN.9
CLK1
CD44.6
...
Expected output
That's fine, but instead what I am after is a 'collapsed' version of this output. In other words, comparing the "transfected" and "non-transfected" subsets, which (other) genes are differentially expressed?
> markers
JCHAIN
CLK1
CD44
...
I am new to these tools, so thank you very much for any assistance and kind guidance in the right direction.
Any thoughts at all..? They might really help out a lot.