Should I use gene_id or gene_name in featureCounts for differential expression gene analysis?

Question

[biock]

Hello! When I use featureCounts to quantify gene expression counts, there are two choices of attribute type in GTF annotation: gene_id and gene_name (specified by -g option). The default value is gene_id, while I found many studies use gene_name. What's the difference between the two choices for downstream differential expression analysis?

Answer 1

[jared.andrews07]

Depends what you're using for the differential expression analysis. Usually used for gene level summarization on data import into DESeq2 or whatever with tximport or the like. So really, it shouldn't matter, but you probably want to use gene_id since those identifiers are generally easier to map to different things.