Question 7 of 20 Fundamentals Concept
When would you choose poly(A) selection over ribosomal RNA depletion for an RNA-seq library?
Short answer
Choose poly(A) selection for intact eukaryotic RNA when the goal is measuring mRNA, because far more reads land on exons. Choose rRNA depletion for degraded samples, for bacteria, or when RNAs without a poly(A) tail matter.
Why
Ribosomal RNA makes up over 90 percent of total RNA, and mRNA only 1 to 2 percent, so a library must either pull out mRNA or remove rRNA. Poly(A) selection captures molecules by their poly(A) tail with oligo-dT. That works well only when the RNA is intact, usually judged by the RNA integrity number (RIN).
If a transcript is broken, only the piece attached to the tail is captured, and coverage piles up at the 3' end. Ribosomal depletion keeps everything that is not rRNA, including pre-mRNA, so many reads fall in introns. In a study of blood and colon samples, the share of usable exonic reads was 71 versus 22 percent in blood and 70 versus 46 percent in colon for poly(A) versus depletion.
Depletion needed about 220 percent and 50 percent more reads to reach the same exonic coverage. So for a gene expression study on good-quality human tissue, poly(A) selection is usually the cheaper route to the same answer. The choice flips for biopsies or archived tissue with poor integrity, for bacterial RNA, which has no poly(A) tail, and when the question involves non-polyadenylated or intronic transcripts, which depletion captures more broadly.
A strong answer covers
- rRNA dominates total RNA, so mRNA must be enriched or rRNA removed
- Poly(A) capture needs intact RNA and gives more exonic reads per sequenced read
- Depletion suits degraded samples, bacteria and non-polyadenylated RNA
- Depletion yields many intronic reads, so it needs more depth for gene counts
Common mistakes
- Using poly(A) selection on degraded clinical RNA and then reading the 3' bias as biology
- Assuming the two protocols give directly comparable expression values and mixing them in one analysis
- Planning the same sequencing depth for a depletion library as for a poly(A) library
The interviewer might ask next
How would you spot from the data alone that a poly(A) library was made from degraded RNA?