What to do when having differenet versions of ensembl IDs?

Question

[Omar Mohamed]

This is a general question, but I want to know the best practice here. Sometimes, when I have an RNA-seq data, the row names is represented as ensembl IDs which is not very meaningful to me. When I try to map the rownames to gene symbols, I got and error that rownames cannot contain duplicated enteries. So, many genes has different versions (Which I don't totally understand how aligners uses multiple versions)

So, what should I do about that? I think If I kept only one of them or the most varying before normalization and clustering check, I would be biasing the analysis as I ignored few counts!

What should I do?

Answer 1

[rpolicastro]

I keep all gene IDs regardless of whether they map to multiple gene names. If I want to display gene names for some visualization or data presentation (such as a list of DEGs) I'll merge the gene names into the matching gene IDs.

[Omar Mohamed]

that's logical so far, but what if one gene of interrest is not matching another version. like one over expressed and other is not or also overexpressed but with different values?

[rpolicastro]

You can report the gene IDs alongside the gene names.

[Omar Mohamed]

Pretty neat idea. Thank you for sharing