How to Remove rRNA From RNA-Seq Libraries: A Method Comparison

Ribosomal RNA makes up the majority of total RNA in most cell types, so removing it before sequencing is a near-universal step in RNA-seq library prep. But "removing rRNA" isn't one single technique, it's a category of methods, and the one you choose directly affects how clean your final data turns out to be.

     

Method 1: probe hybridization

Biotinylated probes complementary to rRNA sequences bind their targets, which are then pulled down using magnetic beads. This works well for well-annotated organisms with fixed, conserved rRNA sequences, but struggles with non-model organisms, degraded RNA, or rRNA variants the probe panel wasn't designed to cover.      

Method 2: enzymatic digestion (RNase H-based)

DNA probes hybridize to rRNA, and RNase H selectively degrades the RNA within the resulting RNA:DNA hybrid. This is efficient for intact, high-quality RNA, but performance drops noticeably with fragmented or low-input samples, since hybridization efficiency depends heavily on RNA integrity to begin with.      

Method 3: poly(A) selection (indirect depletion)

Rather than removing rRNA directly, poly(A) selection enriches for mRNA via its poly-A tail. This is effective for well-preserved eukaryotic mRNA, but it discards non-polyadenylated transcripts entirely, including many long non-coding RNAs and all bacterial RNA, so it isn't really a true depletion method in the strict sense.      

Method 4: LASSO capture (riboLASSO)

A newer approach, used in riboLASSO, uses sequence-specific DNA crosslinker libraries that drive a polymer network to selectively engulf rRNA targets through a phase-change mechanism, independent of enzyme efficiency or RNA integrity. Because it doesn't rely on enzymatic cutting, performance is more consistent across degraded or low-input samples than methods 2 and 3.            

                         

Method	Best for	Weak point
Probe hybridization	Well-annotated organisms	Sequence variants, non-model species
RNase H digestion	Intact, high-quality RNA	Degraded or low-input samples
Poly(A) selection	Eukaryotic mRNA enrichment	Loses non-polyadenylated RNA
riboLASSO capture	Degraded, low-input, diverse samples	Newer, fewer long-term datasets

     

Choosing a method

 For high-quality, well-characterized samples, probe or enzymatic methods work fine and are extensively validated in the literature. For degraded, low-input, or non-model samples, methods less dependent on enzymatic efficiency are worth testing first, since that's specifically where traditional methods tend to underperform.