DE Analysis Between A Subset Seurat Object Versus Original Seurat Object
Entering edit mode
8 weeks ago
Zeddicus • 0

This is my first Biostars post and venture into R and seurat, so I apologize for the naivety of my question and if this has been answered elsewhere. It seems like this should be a common type of analysis question, but I have not been able to find out how to go about it via extensive Googling. Here are the steps I have taken and my question:

  1. Started with a complex tissue dataset which formed 11 low resolution clusters in the "original" seurat object.
  2. Subset the cells in "original" cluster-8 to form a separate "subset-1" seurat object containing 5 low resolution sub-clusters.
  3. Subset the cells in "subset-1" sub-cluster-1 to form a separate "subset-2" seurat object containing 6 high resolution sub-clusters.

How can I perform DE analysis of these "subset-2" sub-clusters compared to the rest of the "original" dataset?

Thanks in advance!

scRNA-seq seurat • 388 views
Entering edit mode

You can add a column to the metadata of the original object and labelled the cells based on the subsets. In pseudocode: ifelse(rownames( %in% cells_subset2,"cluster X","cluster Y").

Entering edit mode

Thanks for your response! I tried:$endothelial_subclusters <- ifelse(rownames( %in% colnames(endothelial_cells), "endothelial_1", "other")

But everything in the new "endothelial_subclusters" metadata column is labelled as "other" and nothing else

Entering edit mode

No worries, I got it to work this way:

Got cell names:

cellnames_all <- Cells(everything)
cellnames_ec <- Cells(endothelial_cells)

Created a vector of labels For cells in the endothelial subset, labelled them as "EC", for others label them as "Other"

labels <- ifelse(cellnames_all %in% cellnames_ec, "EC", "Other")

Added the labels as a new column to the metadata

everything_ec <- AddMetaData(everything, metadata = labels, = "Labels")

Checked if the new column was added

head(, n = 2)

Checked to see if subset of original cells had the right number of labels

Idents(everything_ec) <- "Labels"


Correct number of cells labelled with the "EC" label Can now do perform DE analysis

DefaultAssay(everything_ec) <- "RNA"

ec.vs.all.markers <- FindMarkers(everything_ec, ident.1 = "EC", ident.2 = "Other", logfc.threshold = 1, test.use = "roc", only.pos = TRUE)

Checked list of DE genes

head(ec.vs.all.markers, n = 20)


Login before adding your answer.

Traffic: 2616 users visited in the last hour
Help About
Access RSS

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6