Hello I am using the amazing SingleCellExperiment and SingleR packages to analyze my scRNAseq data. Everything works well and I was able to identify cell types in my samples with the following

pred <- SingleR(test=merged, ref=ref, labels=ref$label.main)

I am looking for a way to look deeper and identify T cell subsets. any advices or suggested tutorials? thank you!

Usually it's a simple matter of swapping ref$label.main for ref$label.fine, which has much finer labels. Obviously, this depends on the reference: BlueprintEncodeData() has a lot of T cell subsets, check out the celldex documentation for more details.

Note that the method will get (a lot) slower when there are many more labels. In particular, this is because SingleR() needs to spend a lot more effort resolving differences between closely-related labels. You can speed it up by throwing more cores at it with BPPARAM=.