rRNA Depletion for Degraded RNA Samples: What Works and Why
Team Dynamic Matrices | 5 October 2026
Degraded RNA is the norm, not the exception: FFPE blocks, biobank samples with long ischaemia times, stored lysates, tissues that are hard to extract. Whether rRNA depletion works on such RNA depends less on the brand than on the mechanism. This guide covers how degradation breaks the common methods, how to measure it, and how we suggest choosing and validating a method.
We make an rRNA depletion kit (riboLASSO). We have tried to keep this vendor-neutral and to point to independent data.
What degradation does to RNA-seq
Fragmentation shortens transcripts; in FFPE tissue most are reduced to fragments of 100 to 300 nucleotides. That has two consequences for rRNA handling. Poly(A) selection fails because the tail is separated from the body of the transcript. And rRNA itself is fragmented, so any depletion method that depends on hybridising to a fixed site, or on a complete enzymatic step, will miss pieces that are too short to be captured or cut. Residual rRNA then consumes reads you cannot afford on a scarce sample.
Why common methods lose efficiency
| Method | What happens on fragmented RNA |
|---|---|
| Poly(A) selection | The tail is clipped from the transcript body, so coverage is confined to the 3' end and yield collapses. Unsuitable for FFPE as a default. |
| Probe hybridisation with bead pulldown | Fragments that do not span a probe site escape, so efficiency drops on FFPE and degraded input (Adiconis et al., 2013). |
| RNase H digestion | Needs stable DNA:RNA hybrids. Short rRNA fragments hybridise poorly and escape digestion. |
| Sequence-selective capture (riboLASSO) | Captures by sequence rather than a fixed probe site, so short fragments carrying any catcher-complementary region can be captured. Validate on your own samples. |
What independent comparisons show
Schuierer et al. (BMC Genomics, 2017) compared three Illumina protocols, poly(A)-based TruSeq Stranded mRNA, TruSeq Ribo-Zero rRNA removal and TruSeq RNA Access exome capture, across inputs from 100 ng down to 1 ng and three levels of degradation. On degraded RNA, the rRNA removal protocol gave more accurate and more reproducible gene expression results than the other two, even at 1 to 2 ng input. On highly degraded RNA, exome capture performed best, giving reliable data down to 5 ng. Cieslik et al. (Genome Research, 2015) likewise showed that exome-capture RNA-seq gives high-quality data from degraded FFPE tissue.
Measure first: DV200, not RIN
RIN relies on intact 18S and 28S peaks, which fragmentation destroys. DV200, the percentage of fragments longer than 200 nucleotides, correlates far better with library success. Our working bands:
| DV200 | Interpretation | Strategy |
|---|---|---|
| Over 70 percent | Good for FFPE | Standard rRNA depletion, whole transcriptome |
| 50 to 70 percent | Usable | rRNA depletion; increase input; expect 3' bias |
| 30 to 50 percent | Compromised | Adjusted protocol; consider targeted capture |
| Under 30 percent | Poor | Targeted capture of predefined genes, or do not sequence |
Measure DV200 on every sample before library prep. It takes minutes and saves failed libraries. More in FFPE RNA extraction challenges and RNA QC beyond A260/A280.
Practical settings for degraded RNA
- Do not default to poly(A) selection. Use rRNA depletion, or targeted capture when DV200 is low.
- Increase input where you can. 50 to 100 ng is a reasonable floor for whole-transcriptome work on FFPE, and a 10 µm section often yields only 10 to 100 ng.
- Check the library prep kit's guidance for fragmented input. Many protocols shorten or skip the fragmentation step when RNA is already short.
- Remove genomic DNA properly. Use a validated, extended DNase step and confirm by no-RT qPCR.
- Run a control block. A block of known quality alongside each batch separates extraction failures from sample failures.
Validate a method on your own samples
Published comparisons use reference RNA; your blocks are different. Before you commit a cohort, take RNA from two or three representative samples, split each, and run the candidate methods side by side. Compare:
- rRNA before and after depletion by qPCR for 18S or 28S, which takes about an hour.
- The fraction of reads that map to rRNA after sequencing.
- Gene detection and expression against an undepleted control, to see whether non-rRNA transcripts are lost.
- Consistency across technical replicates.
Where riboLASSO fits
riboLASSO captures rRNA by sequence with DNA catcher strands on a soluble polymer that condenses into a pellet, rather than relying on a fixed probe site or enzymatic cutting, which is why we designed it for fragmented input. It is released for human total RNA, and we suggest a side-by-side on your own samples before you commit a cohort. See the riboLASSO kit and how it compares with other rRNA depletion kits.
References
- Schuierer S et al. A comprehensive assessment of RNA-seq protocols for degraded and low-quantity samples. BMC Genomics 18 (2017), doi:10.1186/s12864-017-3827-y.
- Adiconis X et al. Comparative analysis of RNA sequencing methods for degraded or low-input samples. Nat Methods 10, 623–629 (2013).
- Cieslik M et al. The use of exome capture RNA-seq for highly degraded RNA with application to clinical cancer sequencing. Genome Res 25, 1372–1381 (2015).
- Morlan JD, Qu K, Sinicropi DV. Selective depletion of rRNA enables whole transcriptome profiling of archival fixed tissue. PLOS ONE 7, e42882 (2012).
- Matsubara T et al. DV200 index for assessing RNA integrity in next-generation sequencing. BioMed Res Int 2020, 9349132 (2020).
- Dynamic Matrices. LASSO: sequence-selective biomolecule isolation by programmable polymer phase separation. Angew Chem Int Ed (2025).
FAQs
There is no single best kit; the mechanism matters. Methods that rely on hybridising to a fixed probe site or on complete enzymatic digestion lose efficiency on short fragments, and poly(A) selection fails because it needs an intact tail. Choose a method validated on fragmented RNA and confirm it on your own samples.
Not as a default. Fragmentation clips the tail from the transcript body, so coverage is confined to the 3' end and yield collapses.
As working bands: above 70 percent is good, 50 to 70 percent is usable with more input, 30 to 50 percent is compromised, and below 30 percent you should consider targeted capture. rRNA depletion is the standard choice down to about DV200 30 percent.
50 to 100 ng is a reasonable floor for whole-transcriptome work. A 10 µm section often yields only 10 to 100 ng, so plan sections accordingly.
Run a qPCR for 18S or 28S against a housekeeping mRNA before and after depletion, then check the fraction of reads mapping to rRNA after sequencing. Compare against an undepleted control to see whether other transcripts were lost.