Why does it make sense to use gene symbols instead of gene ids in enrichment software?

Question

[Aspire]

Hi, newbie's question. Enrichr uses symbol names instead of gene ids for gene set enrichment analysis. Why does it make sense?

Gene symbols can be the same for human and mouse. Shouldn't a person want to distinguish the species s\he is working on, and look at the enrichment at this species only?

P.S. I would also like to ask if gene names can change over time. If I have data annotated with Ensembl 75 release, should I worry about performing an enrichment with Enrichr? In truth, I am not sure what version Enrichr is working with.

[Friederike]

Gene symbols can be the same for human and mouse.

They typically differ sublty, e.g. Tp53 vs. TP53, Gapdh vs. GAPDH etc. Human gene symbols should be all caps.

gene names can change over time

Yes. There are tons of genes with multiple names.

[Aspire]

I did not find it in the documentation, but it seems that Enrichr does not distinguish between human and mouse genes. Enrichr gave me the same GO enrichment results regardless of whether I submitted human genes (all caps) or the same genes for mouse (capitalizing first letter).

Isn't it true that the results should be different if one compares a given gene list to GO annotations of human and of mouse? Why does it make sense not to not make the distinction between human/mouse when performing enrichment?

Answer 1

[igor]

Technically, IDs are better. However, that assumes that the entire analysis pipeline is using IDs. In the real world, pathways are frequently reported only as gene symbols. You may have IDs, but you'll definitely have symbols. Additionally, many pathways were established before any gene IDs even existed. If you are a database aggregator such as Enrichr, it's easier to work with gene symbols.