After read mapping, count exons or genes?

Question

[tianshenbio]

I have mapped my RNA-seq reads (paired-end) onto the reference genome. Now I hope to measure the expression level by genes using featurecounts. Do I count the number of reads mapped to the whole 'gene' region, which includes introns? Or I should count reads mapped to all exons belonging to the same gene? Someone suggested using exons, may I know the reason behand this?

Answer 1

[ATpoint]

Exons because that is what the final spliced transcript is composed of. There is a reason that exon is the default setting in featureCounts.

[tianshenbio]

Ok now I understand, thank you.