I am currently working in Seurat v5 with an integrated dataset (3 control and 3 treatment samples). I merged the samples, and then I ran a standard workflow on the split samples (split by orig.ident):

> dt <- SCTransform(dt, method = "glmGamPoi",  vars.to.regress = c("percent.mt"))
> RunPCA(dt)
> IntegrateLayers(object = dt, method = RPCAIntegration, 
                           normalization.method = "SCT", orig.reduction = "pca", 
                           new.reduction = "integrated.rpca")

I then found the clusters in the integrated object. Now I want to identify the markers with the "FindAllMarkers" function. However, based on the Seurat v5 tutorial, I am a bit confused.

1. Do I have to join the layers again before running the "PrepSCTFindMarkers" and "FindAllMarkers" functions?
2. Does the default assay need to be "SCT"? Also, in the FindAllMarkers function, does the "slot" need to be set to "data" or to "counts", and why? See below:

> DefaultAssay(dt) <- "SCT"
> Int_markers <- FindAllMarkers(Integrated, slot = "data", only.pos = TRUE, 
                              min.pct = 0.25, logfc.threshold = 0.25, test.use = "wilcox")

1. Why in this tutorial (https://satijalab.org/seurat/archive/v4.3/sctransform_v2_vignette) they use the assay = "SCT" in the FindAllMarkers function and is it always required/default?

Thank you so much.