tRNA differential expression analysis in mRNA sequencing data

Question

[KG]

Hi,

I have analysed an mRNA sequencing dataset (~30 million reads for each of the two biological replicates per sample for two samples, generated from poly-A enrichment library) using STAR-Featurecounts-EdgeR protocol. I see 110 tRNA genes are differentially expressed in my mutant compared to control (DE genes defined as log2FC >1.5 or log2FC <-1.5 along with FDR <0.05).

Now, since tRNA transcripts are not poly adenylated (unless marked for degradation), they are not supposed to be enriched in my mRNA sequencing library. But if those tRNA molecules are coming into my library just due to random chance and their relative abundance, is it correct to analyse differential expression for them? The read counts for some of them are very low (0-50) but most of them are > 50.

I would like to know if 1. anybody has seen tRNA reads in mRNA sequencing experiment done with poly-A enrichment library and 2. if yes, then it is correct to compare their expression levels?

Any thoughts?

Thank you for your help!

Answer 1

[Carlo Yague]

Hi, tRNA (and rRNA) contamination of poly-A enriched samples is expected due to their relative abondance indeed. But to compare their expression, you would need to assume that the "contamination" is the same in all your samples which is probably not true.

If you want to make sure that the difference you observe is not an artefact, I recommand to check the expression of a few tRNAs by Northern blot (migration of total RNA on 10% acrylamide gel + hybridation of a radioactive probe). For me this is the best method to study tRNA expression because 1- there is no reverse transcription involved (RT of tRNAs is tricky because of secondary structures and RNA modifications) and 2- it allows you to discriminate between the tRNA isoforms (precursors vs mature tRNAs).

Answer 2

[KG]

Thank you, Carlo! That is an excellent suggestion. I am planning to check for a couple of tRNAs for which the raw read counts are high and have log2FC lower than -4.0. I have done Southern blotting experiments, but not done northern blotting experiments myself. I understand northern blotting procedures will be analogous to Southern blot. It will be very helpful if you could share a working protocol for northern blotting.

[Carlo Yague]

Northern blots are basically the same as southern blots except that you use RNA samples instead of DNA. Therefore, all gels and buffers must be made with RNAse free water and you should wear gloves all the time.

Here is a general protocol for Northern blots. 10% acrylamide gels are the best for tRNAs but be careful not to overload the gel (I never load more than 10 ug/well). For the migration, 225V during 3h at room temperature gives good results but if you need extra-resolution, consider running at 110V during 20h at 4°C as they do here. Good luck.