Entering edit mode
3.5 years ago
barix
▴
20
Hi!
I have single-cell rna seq data and would like to do gene enrichment analysis using DAVID (experimental group vs. control).
- Should I give as an input filtered list (fdr < 0.01) of DEx genes? Or the whole list of ~20,000 genes (unfiltered list)?
- Should I include both, upregulated and downregulated genes?
Thank you!
Generally, I would go for the filtered list. I don't understand why would anyone use the whole list of genes for functional (you can use it as a background for enrichment).
The answer to your second question is purely based on your biological question. If you want to comment on overall DGEs and their biological implication (Pathway enrichment) then of course you have to use all DGE otherwise either go for Upregulated OR Down-regulated genes.
Thank you! It all make sense much more to me now.
If my question is for example how enriched is specific GO term in the list of my DEx genes, would THIS approach be the way to go?
Yep.
fGSEA is an R package and if you are not familiar with programming you can use GSEA GUI desktop version
Thanks! How about enrichment of individual genes (gene by gene) from the list in just one cell type? I'm not sure how to do setup/approach here.
I guess the pre-ranked gene set enrichment analysis option would be helpful in this scenario, which is available in GSEA software.
OR
if you are only interested in the overlap of two gene lists then probably hypergeometric distribution could help.