Any reason for scRNA data of mouse lung samples to fail to capture epithelial cells?

Question

[윤여가(대학원 바이오정보학)]

A project in my lab involves single cell RNA-seq data analysis of mouse whole lung samples. However, when we analyzed and clustered the dataset and searched for markers to annotate the group, there appears to be few to no cells exhibiting the classical epithelial cell markers EPCAM and CDH1. Each sample has around 8000 cells after filtering for mitochondrial content, and the overall quality seems fine. But over half of the cells appear to be immune cells and there were less than 100 epithelial cells for each sample.

The mouse models are established by a collaborating group, and whole lung samples are sent to a sequencing company (travel time about 3 hours minimum) to generate the scRNA data. Our collaborators have adjusted experimental protocols multiple times to increase cell viability (~85%), but we are having difficulty fixing this lack of epithelial cells.

Does anyone have some experience with this, or know why there would be so few epithelial cells in scRNA-seq data of mouse lung samples?

Both my lab and our collaborators are fairly new to handling scRNA-seq data, so any insight would be helpful.

Answer 1

[Matthias Zepper]

That is indeed odd, one would expect plenty of epithelial cells. Did you already check if the lung sample in the Tabula Muris dataset has a similar composition?

As far as the wet lab is concerned, I would compare your experimental procedures with those published at Protocols.io (#1, #2) or in the Tabula muris paper, if there have been any obvious differences. In particular, the protocols from the Human Cell Atlas Method Development Community have been thoroughly vetted and should work well.