Why do Sox9 and Amh, which are Sertoli cell markers, show expression in the overall cell population?

Question

[Bio]

Hello, I am currently analyzing testis-specific scRNA-seq data. I used a cluster resolution of 0.15 to identify the number of clusters and attempted to annotate cell types using known marker genes. However, I noticed that Sox9 and Amh (markers for Sertoli cells) show expression across most cell populations. Specifically: Sox9 is highly expressed in Cluster 0 (log2FC = 1.39) but also shows lower expression in Cluster 7. Amh has higher expression in Clusters 5 and 7.

Questions:

1: Why do Sox9 and Amh, which are Sertoli cell markers, show expression across multiple cell populations?

2: Could there be an issue with my normalization method?

NormalizeData(seurat_obj, normalization.method = "LogNormalize", scale.factor = 10000)

3: Could background noise still be present?

4: I already removed doublets using DoubletFinder, but my input consists of a single sample (a mixture of two days of data).

5: Is it possible to adjust the shape of the UMAP plot?

If anyone of you have any suggestions or insights, I would greatly appreciate your help. Thank you!

Answer 1

[ATpoint]

1: Why do Sox9 and Amh, which are Sertoli cell markers, show expression across multiple cell populations?

Markers are often just enriched, not specific. Depends highly on the marker and celltype(s). The area around x=y=0 looks like it could qualify for this gene, no?

2: Could there be an issue with my normalization method?

It's just scaling normalization, I don't see how for top markers it would matter.

3: Could background noise still be present?

In single-cell, and biology, and science, yes, always. (corny answer I know, but yes, noise can always be an issue).

4: I already removed doublets using DoubletFinder, but my input consists of a single sample (a mixture of two days of data).

What's the question?

5: Is it possible to adjust the shape of the UMAP plot?

Yes, check Seurat docs on how to do that. In the underlying uwot::umap() which I guess it uses you have options like spread and min.dist.

As for markers, if just two markers are not enough then see whether you can fetch a signature for your celltype and score the cells against it. See if a cluster shows enrichment for it. It's open-ended, so I cannot give a definite answer.

[Bio]

Thank you for your detailed response—it was very helpful! I’ve checked additional markers (like WT1), but they still show a similar expression pattern. Could this suggest my results are correct, or might there be technical/biological noise influencing this? I’d greatly appreciate your thoughts on how to interpret or validate this further.

[ATpoint]

I know nothing about these cells or the experiment, I really cannot tell. You have to ask a technical question that one can actually answer without knowing your biological or experimental context.

[Bio]

Thank You for your time. I have single-sample data from two developmental stages of mouse testis, and we need to identify the transitional stages between them. Since the testis contains both germ cells and somatic cells, we must assess the expression of specific marker genes. However, I observed that the Sertoli cell marker genes (Amh, Sox9, WT1) were expressed across the entire cell population—a finding I already shared with you in the figures.

[ATpoint]

You have to ask a technical question that one can actually answer without knowing your biological or experimental context.

[Bio]

Thank You for your time and answered me in details