Challenges with RNA Extraction from FFPE Samples: Why Formalin-Fixed Tissue Keeps Breaking Your RNA-Seq Workflow

Team Dynamic Matrices | 28 July 2026

Formalin-fixed, paraffin-embedded tissue represents decades of clinical history. Tumor banks, biobanks, and pathology archives worldwide hold millions of blocks, each one a potential source of transcriptomic insight into disease progression, treatment response, and biomarker discovery. The problem is that FFPE tissue was never prepared with RNA sequencing in mind. It was prepared to preserve morphology for a pathologist looking through a microscope, and the fixation chemistry that makes tissue architecture look pristine for decades is the same chemistry that shreds and chemically modifies RNA. Anyone who has tried to pull a clean, full-length RNA-seq library out of an FFPE block knows this tension firsthand.

What formalin actually does to RNA

Formaldehyde fixation works by forming methylene bridges between nucleophilic groups on proteins and nucleic acids. This crosslinking is exactly what locks tissue structure in place, but it also covalently traps RNA in ways that are difficult to reverse. During long-term storage, RNA in FFPE blocks undergoes progressive fragmentation, with most transcripts reduced to fragments in the 100 to 300 nucleotide range within a few years. On top of fragmentation, formaldehyde adducts (mostly mono-methylol modifications on adenine, guanine, and cytosine bases) accumulate over time and interfere with reverse transcription, causing polymerase stalling, misincorporation, and truncated cDNA.

The net effect is a sample that is simultaneously short, chemically modified, and often present in vanishingly small quantities, particularly when the block is a needle biopsy or a small resection from years ago. Any extraction and library prep strategy has to survive all three problems at once, not just one of them.

Why standard extraction kits struggle

Most silica-column or bead-based total RNA extraction kits were optimized for fresh-frozen tissue or cultured cells, where RNA is intact and abundant. Applied to FFPE, several failure modes show up consistently.

Deparaffinization and heat-based decrosslinking steps are a balancing act. Insufficient heat leaves crosslinks intact and RNA unavailable to enzymes downstream. Excessive heat or overly aggressive proteinase K digestion accelerates further RNA fragmentation, so labs often end up trading one problem for another rather than solving both.

Genomic DNA carryover is also worse in FFPE than in fresh tissue, because the same crosslinking that traps RNA also traps DNA fragments in overlapping size ranges, and DNase treatment on already-fragmented RNA introduces additional degradation risk with each handling step.

Perhaps most consequentially, standard total RNA extraction does nothing to address the sequencing problem that follows extraction. Even a technically clean FFPE RNA extract is still dominated by degraded ribosomal RNA and fragmented, formaldehyde-modified transcripts. Poly-A selection, which depends on intact 3' ends and full-length poly-A tails, performs poorly on FFPE RNA because so many of the poly-A tails have already been clipped by fragmentation. This is why FFPE RNA-seq libraries prepared with standard poly-A workflows so often show 3' bias, low library complexity, and poor gene body coverage, even when the input extract passed every quality check.

Quality control that actually predicts sequencing success

A260/A280 and A260/A230 ratios tell you about protein and chemical contamination, but they say almost nothing about fragmentation or formaldehyde adduct load, the two variables that most determine whether FFPE RNA will sequence well. The DV200 metric, which reports the percentage of RNA fragments longer than 200 nucleotides, has become the standard proxy for FFPE RNA quality because it correlates far better with library prep success than absorbance ratios or even RIN scores (which were designed around intact 18S and 28S ribosomal peaks that FFPE samples rarely retain).

A DV200 above 50 percent is generally considered workable for standard library prep, while samples between 30 and 50 percent require adjusted input amounts and more forgiving chemistry. Below 30 percent, most poly-A and even many exome-capture workflows produce libraries too sparse and too biased to interpret confidently. Knowing where a sample falls before committing sequencing budget to it is essential, but knowing the number does not by itself fix the underlying problem: short, modified, low-abundance transcripts still need a method that works with those constraints rather than against them.

Why target capture changes the equation

Ribo-depletion helps by removing the dominant ribosomal RNA fraction and does not depend on intact poly-A tails, which is why it has largely replaced poly-A selection for FFPE work. But depletion alone still leaves you sequencing whatever fragments happen to be in the tube, informative or not, and on severely degraded or vanishingly low input samples, the signal of interest can remain diluted below a useful detection threshold.

Target capture takes a different approach. Rather than removing unwanted RNA and hoping enough target transcript survives, sequence-specific probes pull down only the transcripts of interest, and everything else is washed away. This matters enormously for FFPE, because it means the sequencing budget is spent on the regions of the transcriptome that answer the biological question, not on whatever fragments of degraded housekeeping transcripts happened to survive fixation best.

LASSO's target capture is built on a swollen polymer network, not a bead surface or a simple solution-phase probe format, and that structural choice is what makes it well suited to degraded FFPE input. The swollen network is what gives the chemistry its selectivity, retaining true target sequences while excluding near-identical off-target fragments more cleanly than probes alone. Because the network presents a high surface-to-volume ratio, binding capacity stays high even when starting material is scarce and fragmented, which is exactly the situation a small FFPE biopsy or an old archival block creates. And because the capture matrix is built from polymer materials rather than specialized bead chemistries, the cost per sample stays low, which matters when a lab is enriching across large retrospective cohorts rather than a handful of samples.

This becomes particularly valuable in a few recurring scenarios. Retrospective cohort studies often need transcript-level answers from blocks that are five, ten, or twenty years old, where DV200 values are low and total RNA yield is minimal. Needle biopsies and other small clinical specimens frequently do not generate enough intact RNA for whole-transcriptome sequencing at any reasonable depth, but a targeted panel focused on a defined gene set can still deliver quantitative, reproducible results from the same limited input. And in translational research where a specific pathway or biomarker panel is the actual endpoint, targeted capture avoids spending sequencing depth on genome-wide noise that will never be analyzed.

Practical recommendations for FFPE RNA workflows

A few principles consistently improve outcomes when working with FFPE tissue. Extraction protocols should be matched to block age and fixation time where that information is available, since longer formalin exposure generally means more crosslinking and shorter fragments. DV200 should be measured before committing to a library prep strategy, not after a failed sequencing run reveals the problem retroactively. Ribo-depletion or targeted capture should be the default choice over poly-A selection for any FFPE input, given how unreliable intact poly-A tails are in these samples. And where the biological question is defined in advance (a specific pathway, a validated biomarker set, a known gene panel), targeted enrichment should be considered before whole-transcriptome sequencing, both for cost efficiency and for signal quality on difficult samples.

FFPE tissue will keep being the material most researchers actually have access to, long after fresh-frozen collection becomes standard practice going forward. The extraction and library prep strategy has to be designed around what FFPE RNA actually is: short, chemically modified, and often scarce, rather than treated as a lower-quality substitute for fresh-frozen input that just needs a more careful pipette hand. Matching the method to the material, rather than forcing the material through a workflow built for something else, is what turns archival blocks back into usable transcriptomic data.