Why does biomart return multiple Entrez IDs

Question

[hdy]

I am using biomart to convert multiple gene names to Entrez IDs, however, I found that for some gene names, it returned multiple IDs. Shouldn't the name and ID have a one to one correspondence? I also tried to use all the returned IDs and search in NCBI, but it seems only one of them is the same with my inquiry. For example, try "POTED" in biomart, it returns three IDs - "388468,339010,317754". But only searching 317754 in NCBI will give you the right gene symbol POTED. Why biomart returns multiple IDs?

Answer 1

[Devon Ryan]

Gene names aren't unique. The closest to an actually unique ID that I normally use would be Ensembl gene IDs (this is why I normally recommend using Ensembl annotations, when available).

Edit: I'll add that a single Entrez ID can have multiple gene names (on more than one chromosome or even on different strands of the same chromosome). Yeah, that can make life difficult.

[Arindam Ghosh]

In RNA-seq data analysis do we merge counts of such genes if multiple Ensembl gene IDs have same gene name or treat it separately. I recently encountered this issue. All was good till I was working with Ensembl gene IDs. Just when I have to annotate it I find the gene names were same. How do you deal if such genes come as differentially expressed or network hubs?