Entering edit mode
3.2 years ago
Dear community,
I want to ask a question regarding the presence of histone and lncRNAs genes in poly(A) RNA-seq data. I know that some of them are polyadenylated but I guess it is more an exception than the rule.
I wonder if there are any recommendations on how to deal with it, for instance, in tasks related to counts table preparation and subsequent differential expression analysis. Is it better to remove or ignore?
Thank you in advance, Best wishes, Alexandr
What is a histone RNA? You mean genes encoding histone proteins?
That would be a normal protein-coding RNA.(Edit: Stupid me did not know many histone genes lack polyA tails, sorry). Many mncRNAs are polyA, so yes you will see them, and I see no reason to exclude them. Just use a GTF reference file and quantify against it. No magic here. Is there a specific concern you have?Thank you for your answer!
Sorry for inaccuracy in terms. I mean mRNA of genes coding histones.. they are not polyadenilated as I know (but not all). So, it is unexpected for me to observed read counts releted to this genes in poly A RNA-seq data.
Thanks!
Oh sorry, I did not know that toward histone mRNAs.
Still, I cannot remember that I ever heard people filtering histone mRNAs. The thing is that when you start bothering with these kinds of things the filtering does not stop. You would probably want to remove all the small RNAs, and probably the genes not experimentally confirmed (TECs), and maybe pseudogenes and...and...and. Therefore, to keep it simply and standardized, just use all genes that are in the reference you use. That is what probably 99.5% of people do, otherwise filtering is very opinion-based and therefore difficult to justify. Others might have better suggestions, but I would really not overthink it.
Thank you very much for your answer!
I always did my analysis as you described. But now I have Ribo-seq data coupled with poly(A) RNA-seq. And some histone genes are presented in the set of genes with differential translation efficiency (the set of interest for me) and thus could impact downstream analysis like GSEA and GO clusterization.
But as you said, I too have never met a recommendation to filter them from the analysis.
Best wishes, Alex