Test custom signature using GSEA in scRNA-seq

Question

[bio_info]

Hello all! I am trying to carry out GSEA analysis and run ClusterProfiler in R to test a custom gene signature (comprising of 7 genes) in my scRNA-seq dataset, on the output of FindMarkers. In this case, I think the background will be the DEG and the 'test' will be my custom gene signature. I have processed the dataset using Seurat and I am a bit confused about the workflow and steps I need to follow.

I have followed several tutorials online, but I would be grateful if someone could help answer these questions:

1. Do I need to modify the output of FindMarkers in a way that it is accepted by ClusterProfiler? How do I go about this?

2. Do I need to carry out GO pathway analysis after running FindMarkers but before running ClusterProfiler?

If possible, could someone please share example code for these steps. I am quite confused and would appreciate all the help!

Answer 1

[ATpoint]

What you describe is an overrepresentation analysis, enricher from the package can do that. Check its documentation. Unclear though what insight you could get from just 7 genes.

[bio_info]

Thank you for your response! The 7 genes are a specific signature for a cell-type I have been asked to use. Would I also need pathway information for using enricher and visualizing the output of GSEA?

[ATpoint]

What exactly do you want to analyze? I understand it as if sou want to see whether this signature is enriched for something. aor do you want to test which clusters are enriched for it? If the latter I would probably just plot signature scores (UCell package) per cluster.

[bio_info]

I want to check the former: Whether my custom signature is enriched or not in the output of DEG and a way to visualize this output. I'm not sure if the visualization requires GO analysis...