Differentially expressed genes between Seurat objects

Question

[lic4010]

Hi all,

I'm trying to find differentially expressed genes between two different subsets of a Seurat object. Is there a straightforward way to do this using FindMarkers() or FindVariableFeatures()? The subsets are pre- and post-treatment and my goal is to find and remove the genes that are associated with the treatment.

Thanks!!

[ATpoint]

Do you have biological replicates? I am personally a fan of pseudbulk DE for single-cell data. In my hands this denoises the issues related to sparse data quite a lot when it comes to DE testing.

Answer 1

[jared.andrews07]

Keep the samples in the same Seurat object, use a model-based test, and include the condition as a latent variable in FindMarkers, i.e. test.use = "LR", latent.vars = "treatment".

[rpolicastro]

I just want to add that you should use the counts from the RNA assay instead of other assays like SCT or integrated.