3' mRNAseq Bioinformatic Considerations (bulk RNAseq)

Question

[stoptime_armstrong]

I am having difficulty finding materials outside of product sheets concerning bulk RNAseq using polyA tail selection (3' transcriptomics/ 3'mRNAseq)

Any materials or papers that cover metrics unique to 3' RNAseq (bulk samples not looking at single cells) will be appreciated.

So far I have gleaned that counting reads using a GTF file containing 3' UTRs is vital.

What are other considerations I should make? I already found examples of pipelines for these kind of data.

So I am more concerned with QC analysis.

I imagine there will be metrics for 3' RNAseq that will be different than random fragmentation of transcripts. So any help in sourcing these will be appreciated. Thank you.

[Shred]

Did you used QuantSeq from Lexogen for library preparation? I've worked with this technology, may be so painful to analyze. Which organism?

Answer 1

[Carlo Yague]

Many interesting considerations are discussed in this paper from Bin Tian group. In brief, what's important from my perspective:

• mispriming events from oligo-A internal stretches is a common issue in polyA-seq (need to be aware of this bias). Some lib prep methods are more robusts than other against this bias.
• alternative poly-A sites (PAS) can be found outside 3'UTRs
• A typical metric used in such studies is the PAS count: the number of reads ending at one position, which indicates the cleavage and poly-adenylation sites on the mRNA/ncRNA. Cleavage is a stochastic process, so it is best to consider small windows as discrete unit of analysis.
• When multiple PAS are found on one gene, we have alternative cleavage and poly(A) (APA). The proportion of proximal PAS vs distal PAS usage (pPAS vs dPAS) can change in different conditions.
• A useful control is that major PAS are typically preceded by the AAUAAA consensus sequence about 17 nucleotide sequences upstream the cleavage site.