The concept of differentially expressed genes at exon and intron level

Question

[Apex92]

Dear all,

I was thinking about the concepts below - I would appreciate your input.

A. Can differentially expressed genes at the intron level (counts based on reads that map to intronic regions) exactly determine transcriptional activity?

B. If a given gene due to a particular regulator is highly down-regulated at the exon level but slightly downregulated at the intron level, can that mean that the effect of that particular regulator on that gene is higher at the post-transcriptional level compared to the transcriptional level?

Thank you in advance.

Answer 1

[Istvan Albert]

Traditional RNA-Seq sequencing operates on mature mRNA meaning that you won't get information for the intronic regions so you can't

That being said there are methods for quantifying pre-mRNA as described here:

• Tracking pre-mRNA maturation across subcellular compartments identifies developmental gene regulation through intron retention and nuclear anchoring Genome Research, 2021

the paper may provide you with some answers

[Apex92]

Dear Albert - thank you for your answer. I agree with you but how about having a custom gtf file with coordinates for introns and then counting genes based on that?

[lieven.sterck]

Won't help.

What Istvan is saying (and that is indeed how it all works) is that the biological molecule that you sample to have then sequences (and thus from which the RNAseq reads are derived) is the mature mRNA, hence those do not contain intron sequences anymore (they have been spliced out already) and you will thus not have them present in your read pool.

This is the theory of course. In practise it can happen that you still get (some) intronic sequences as for instance an mRNA might not have been fully spliced yet at the time of sampling but those will be only a marginal fraction of your total amount of reads.

there are also other biological reasons which can lead to intron sequences being captured in your RNAseq, for instance alternative splicing (and then mainly the intron-retention kind), but again seen in the totality of your reads this will only be a very small fraction