Stereo-seq Sample Preparation and QC: Fresh Frozen vs FFPE Tissues

Inquiry

Figure 1. Overview of the two Stereo-seq sample preparation paths — fresh frozen (left) and FFPE (right) — from tissue collection to chip-ready section.

Sample preparation is the single largest source of variability in Stereo-seq projects. A well-prepared tissue section preserves RNA integrity, maintains spatial architecture, and produces clean spatial gene expression maps. A poorly prepared one wastes sequencing depth on degraded transcripts and blurred spatial signal. This article walks through the sample requirements for both Stereo-seq workflows — fresh frozen (polyA capture) and FFPE (Stereo-seq V2 random-primer capture) — and provides a pre-submission QC checklist you can use before tissue touches a chip.

Key Takeaways

• Stereo-seq runs two distinct sample paths. Fresh frozen tissue uses poly(A) capture on the established V1 workflow and requires tissue-specific permeabilization optimization. FFPE tissue uses the newer V2 random-primer workflow, which captures total RNA and skips permeabilization optimization.
• RNA quality thresholds are workflow-specific, not universal. Fresh frozen samples need RIN ≥ 7 for optimal results (RIN ≥ 4 is the floor). FFPE samples need DV200 ≥ 30%, though the V2 workflow has produced usable data from blocks with DV200 as low as 18%.
• Permeabilization optimization is mandatory for fresh frozen. Every new tissue type needs a 4-point time course on a test chip before the formal experiment — skipping this step is the most common cause of failed Stereo-seq runs.
• Section thickness and mounting differ by workflow. Fresh frozen: 10 μm cryosections, thaw-mounted onto the chip. FFPE: 5 μm microtome sections, dried at 42°C before deparaffinization.
• Tissue architecture QC matters as much as RNA QC. DAPI nuclear morphology and H&E structural assessment catch problems that RIN and DV200 miss — necrotic regions, freezing artifacts, and sectioning damage all compromise spatial data quality.

Two Workflows, Two Sample Paths

Stereo-seq does not have a single sample preparation protocol. The workflow splits at the first decision point: the preservation state of the tissue. Fresh frozen tissue follows the established Stereo-seq V1 poly(A)-capture workflow. FFPE tissue follows the newer Stereo-seq V2 (sometimes called OMNI) random-primer workflow. The two paths differ in capture chemistry, RNA targets, QC metrics, sectioning requirements, and whether permeabilization optimization is needed.

The choice between these two paths is usually made for you by sample availability. If you have archived FFPE blocks, the V2 workflow is your only option — and it is now a viable one. If you are collecting samples prospectively and your question centers on mRNA-level gene expression, fresh frozen remains the more established choice with deeper published validation and a broader user community. For a broader discussion of when Stereo-seq fits a project in the first place — across resolution, tissue size, and platform comparisons — see the Stereo-seq platform decision guide.

Fresh Frozen Tissue Preparation

The fresh frozen workflow starts at the point of tissue collection and every step between collection and the chip affects the data quality at the other end.

Collection and freezing. Tissue should be collected and processed as quickly as possible after excision. Rinse the sample gently with 1× PBS or sterile saline to remove blood and debris. Blot dry with lint-free wipes — excess liquid creates ice crystal artifacts during freezing. For freezing, the recommended method is liquid-nitrogen-cooled isopentane: pre-cool isopentane in a liquid nitrogen bath, then immerse the tissue for 1–2 minutes until fully frozen. Direct liquid nitrogen immersion creates gas pockets around warm tissue (the Leidenfrost effect), producing uneven freezing and tissue cracking — avoid it.

OCT embedding. Pre-chill OCT compound on wet ice. Fill an embedding mold with a layer of pre-chilled OCT, position the frozen tissue, and cover completely with OCT. Avoid air bubbles — they create voids in sections. Freeze the block by placing the mold on a pre-chilled metal block on dry ice, or by brief immersion in the liquid-nitrogen-cooled isopentane bath. Store embedded blocks at -80°C.

Cryosectioning. Set the cryostat chamber to -18°C to -22°C (standard tissues) or -25°C to -30°C (fatty tissues). Allow the OCT block to equilibrate to chamber temperature for 30 minutes before sectioning. Cut sections at 10 μm thickness. Discard the first 2–3 sections to clear the block face of OCT-only layers and surface artifacts. Mount each section directly onto the Stereo-seq chip by brief thaw-warming — the section adheres as it warms momentarily against the chip surface. Dry the chip-mounted section on a 37°C hot plate for approximately 4 minutes.

Methanol fixation. Immediately after drying, immerse the chip in -20°C methanol for 30 minutes. This fixes the tissue and permeabilizes cell membranes. After fixation, the chip proceeds to staining and imaging.

Permeabilization optimization. This is the step that first-time users most often skip — and the one that most often causes failed runs. Fresh frozen Stereo-seq requires tissue-specific permeabilization because different tissue types release mRNA at different rates. The optimization uses a separate permeabilization test chip (Chip P) with 4 consecutive sections from the same block. Each section receives a different permeabilization time — typically 6, 12, 18, and 30 minutes with PR enzyme in 0.01N HCl at pH 2.0 — followed by fluorescent reverse transcription (TRITC-labeled nucleotides). After tissue removal, the 4 chips are imaged in the TRITC channel. The optimal time point is the one with the strongest fluorescent signal and the sharpest tissue footprint — strong signal with minimal lateral diffusion. Published optimal times for common tissues provide a starting point:

These times are starting points, not guarantees. Tissue-specific factors — age, fixation quality, cellular density — shift the optimum. Always run the optimization on your specific tissue type and batch before committing experimental sections to the formal capture chip.

FFPE Tissue Preparation With Stereo-seq V2

The Stereo-seq V2 FFPE workflow was published in Cell in mid-2025 and represents a break from the fresh frozen protocol in almost every respect: capture chemistry, RNA species recovered, sectioning specifications, and permeabilization approach. For researchers with archived clinical blocks or pathology collections, it opens a path to spatial transcriptomics that did not exist for Stereo-seq before.

Block selection and sectioning. Select FFPE blocks with documented fixation conditions — ideally, fixation in 10% neutral buffered formalin for <24 hours, which preserves higher DV200 values. Blocks stored at 4°C after embedding are preferred; blocks up to 3 years old are standard, and the V2 study demonstrated usable data from blocks up to 9 years old. Cut sections at 5 μm on a microtome with a fresh blade. As with frozen sections, discard the first 1–2 sections to clear the block face. Float sections on a water bath at approximately 38°C (lower for fatty tissues) and mount onto the Stereo-seq N chip.

Drying. Dry chip-mounted FFPE sections stepwise: 42°C for 3 hours, then 37°C for 16 hours (overnight), then 60°C for 1 hour. This graduated drying schedule minimizes section cracking and detachment. After drying, sections can be stored in a sealed desiccator at room temperature for up to several weeks before processing.

Deparaffinization and decrosslinking. Immerse the dried chip in Sub-X or xylene for deparaffinization (2 × 10 minutes), followed by rehydration through a graded ethanol series (100%, 95%, 70%, 50%, 3 minutes each) to nuclease-free water. Decrosslinking is then performed at 85°C for 30 minutes in a Tris-based buffer. This heat-induced reversal of formalin crosslinks is the critical step that makes FFPE RNA accessible to random primers.

Post-decrosslinking fixation. After decrosslinking, the tissue is fixed in -20°C methanol for 20 minutes. From this point forward, the workflow converges with the fresh frozen protocol: staining, imaging, reverse transcription (using random primers instead of poly(T) primers), cDNA amplification, library preparation, and MGI-platform sequencing.

No permeabilization optimization. A practical advantage of the V2 workflow is that permeabilization is pre-optimized — the random-primer chemistry is applied at a fixed 30-minute incubation without the tissue-specific time course required for fresh frozen samples. This removes one of the more labor-intensive steps from the workflow and makes FFPE Stereo-seq more accessible to labs without prior optimization experience.

RNA Quality Metrics That Matter

Two numbers dominate Stereo-seq sample QC, and confusing them leads to rejected samples or, worse, samples that pass QC but fail in the lab.

RIN — the fresh frozen metric. The RNA Integrity Number (RIN) is measured by capillary electrophoresis (BioAnalyzer or Fragment Analyzer) and yields a value from 1 (fully degraded) to 10 (fully intact). For Stereo-seq fresh frozen samples, RIN ≥ 7 is the target. RIN between 4 and 7 is marginal — it may work, but expect reduced transcript detection sensitivity and potentially uneven spatial coverage. RIN below 4 is generally not recommended. RIN is measured on RNA extracted from a test cryosection or from a tissue lysate prepared in parallel to sectioning.

What degrades RIN:

• Slow tissue collection and delayed freezing
• Repeated freeze-thaw cycles of the OCT block
• RNase contamination during sectioning
• Prolonged storage at -80°C (degradation is slow but cumulative over years)

DV200 — the FFPE metric. DV200 measures the percentage of RNA fragments longer than 200 nucleotides. It acknowledges the reality of FFPE RNA: formalin fixation and paraffin embedding fragment RNA, so "intact" RNA is not the right expectation. Instead, DV200 asks whether there are enough reasonably long fragments for meaningful capture and sequencing. For Stereo-seq V2 FFPE, DV200 ≥ 30% is the standard threshold. However, the V2 workflow uses random hexamer primers rather than poly(T) selection, which makes it more tolerant of fragmentation than poly(A)-dependent methods. In the 2025 Cell study, samples with DV200 as low as 18% produced interpretable spatial data. The practical interpretation: below DV200 30%, data quality declines but may still be usable for exploratory analysis. Below DV200 15%, the risk of failure increases substantially.

Beyond the numbers. Both RIN and DV200 are extracted-RNA metrics — they measure what happens in a homogenized tissue lysate, not what happens at a specific coordinate on the chip. Two samples with identical RIN values can produce different spatial data quality if one has localized RNA degradation around a necrotic core or freezing artifact. This is why DAPI or ssDNA nuclear staining and H&E morphological assessment are not optional add-ons — they are co-equal QC steps. Nuclear staining that appears punctate and well-defined across the section is a strong positive sign. Diffuse, washed-out staining suggests degraded nucleic acid that RIN alone may not catch.

Figure 2. RNA quality assessment decision flow for Stereo-seq — matching the right metric to sample type and interpreting borderline values.

Tissue-Specific Handling and Special Cases

Some tissues are straightforward. Others require protocol adjustments that make the difference between a usable dataset and an unusable one.

Adipose and fatty tissues. Breast tissue, adipose-rich tumors, and liver contain lipids that interfere with OCT embedding and section adherence. Lower the cryostat temperature to -25°C to -30°C. Increase section thickness to 12–15 μm for fresh frozen. Use a lower water bath temperature (~38°C) for FFPE floating sections. Expect softer tissue morphology — lipid vacuoles may appear as empty spaces in the final image, which is normal, not an artifact.

Brain tissue. Brain is lipid-rich and structurally heterogeneous — white matter tracts and gray matter nuclei have different sectioning properties. Standard cryostat settings (-18°C to -20°C) work for most brain regions, but cerebellum and brainstem can benefit from slightly colder temperatures. The main concern with Stereo-seq on brain is lateral diffusion: the 2024 Nature Methods benchmark found stronger diffusion in olfactory bulb and hippocampus compared to denser tissues. A permeabilization optimization that prioritizes sharp signal boundaries over maximum signal intensity is advisable for brain sections.

Calcified and mineralized tissues. Bone, tooth, and calcified tumors require decalcification before sectioning. Standard EDTA decalcification, if prolonged, degrades RNA. Use rapid decalcification protocols (EDTA-based, <24 hours) and verify RIN or DV200 after decalcification before committing tissue to the Stereo-seq workflow. If decalcification times exceed 24 hours, DV200 values tend to drop below usable thresholds.

Fibrotic and collagen-rich tissues. Fibrotic liver, cardiac scar tissue, and desmoplastic tumors section unevenly — dense collagen bundles resist the blade and can cause chatter artifacts (periodic thickness variations). A fresh, sharp blade is essential. Slower sectioning speeds and slightly thicker sections (12 μm for fresh frozen) reduce chatter. Collagen autofluorescence in imaging channels does not affect Stereo-seq capture chemistry but can complicate the staining and imaging QC step — consider ssDNA staining (Qubit ssDNA reagent) rather than H&E for fibrotic samples.

When to run a pilot. If your tissue type is not listed above and has no published Stereo-seq data, run a single pilot section through the full workflow before committing the full sample set. A pilot tells you whether the permeabilization time range is appropriate (fresh frozen), whether the tissue adheres to the chip through all wash steps, and whether RNA recovery is sufficient for your biological question. The cost of a pilot chip is small relative to the cost of redoing an entire experiment.

Pre-Submission QC Checklist

Working through this checklist before submitting samples catches the issues that degrade spatial data quality. Each item corresponds to a common failure point observed across Stereo-seq service projects.

RNA and nucleic acid integrity

• Fresh frozen: RIN measured and ≥ 7 (or ≥ 4 with acknowledged risk)
• FFPE: DV200 measured and ≥ 30% (or ≥ 18% with acknowledged risk)
• RNA concentration measured (Qubit or Nanodrop) — sufficient for library preparation
• RNA purity: A260/A280 ratio between 1.8 and 2.1 (Nanodrop)

Histology and tissue architecture

• Adjacent section stained with H&E or ssDNA — tissue morphology confirmed intact
• No evidence of necrosis covering >30% of the region of interest
• No freezing artifacts (ice crystal voids) in fresh frozen sections
• No section folding, tearing, or detachment from the slide surface
• Nuclear staining (DAPI/ssDNA) appears punctate and well-defined across the section

Section quality and documentation

• Section thickness confirmed: 10 μm (fresh frozen) or 5 μm (FFPE)
• Section placed within the active capture area of the chip (no overhang)
• No air bubbles or OCT residue visible on the section
• Section ID, block ID, and orientation documented with a photograph of the chip
• For fresh frozen: permeabilization optimization completed and optimal time recorded

Shipping and handling

• Fresh frozen: shipped on sufficient dry ice (≥5 kg for overnight); block wrapped in parafilm to prevent dehydration
• FFPE blocks: shipped at 4°C with cold packs; FFPE sections: shipped at room temperature in sealed slide mailers with desiccant
• Shipping manifest includes block/section IDs, tissue type, preservation date, and RIN/DV200 values

For researchers preparing samples for Stereo-seq through a service provider, the FFPE spatial transcriptomics service page covers additional sample submission requirements and feasibility review steps. General spatial transcriptomics service information, including sample intake and project scoping, is available through CD Genomics Spatial Transcriptomics Services.

Figure 3. Pre-submission QC checklist for Stereo-seq samples — six categories to verify before tissue reaches the chip.

FAQ

Q: Can I use Stereo-seq on samples that were originally collected for scRNA-seq?

If the samples were fresh-frozen without OCT embedding, they can be embedded in OCT and sectioned for Stereo-seq — but RNA quality degrades with each freeze-thaw cycle, so measure RIN first. Samples already dissociated into single-cell suspensions cannot be used: Stereo-seq requires intact tissue architecture. If the samples were cryopreserved in DMSO or other cryoprotectants, wash the tissue briefly in PBS before OCT embedding, and be aware that cryoprotectant residues can affect permeabilization kinetics.

Q: What if my FFPE block is older than 3 years?

The Stereo-seq V2 study tested blocks up to 9 years old and obtained usable data. The key variable is DV200, not age. Measure DV200 from a 20 μm curl or a single 5 μm section from the block. If DV200 ≥ 30%, proceed with standard expectations. If DV200 is 18–30%, the V2 workflow may still work — consider running a pilot section first. If DV200 is below 18%, the risk of failure is high, but the random-primer chemistry gives Stereo-seq V2 more tolerance for fragmentation than poly(A)-based spatial methods.

Q: Do I need to run permeabilization optimization for every FFPE block?

No. The Stereo-seq V2 FFPE workflow uses a fixed permeabilization protocol — no tissue-specific optimization is required. This is one of the main practical advantages of the V2 chemistry. Fresh frozen samples, however, do require tissue-specific optimization for every new tissue type.

Q: How many sections should I prepare for one Stereo-seq run?

For fresh frozen: prepare at least 5–6 consecutive 10 μm sections — 4 for permeabilization optimization (Chip P) and 1–2 for the formal capture chip (Chip T), plus 1–2 spare sections in case of mounting failure. For FFPE: prepare 4–6 sections at 5 μm — 1–2 for the formal capture chip, plus spare sections and at least one for DV200 measurement. Always cut more sections than you think you need: mounting failures, tissue detachment during processing, and staining artifacts are all more common with spatial transcriptomics chips than with standard glass slides.

Q: What is the most common reason for Stereo-seq sample rejection at service providers?

Insufficient RNA quality — specifically, RIN below the stated threshold for fresh frozen samples, or DV200 below 30% for FFPE samples — accounts for the largest share of rejections. The second most common reason is tissue morphology issues: necrosis, extensive freezing artifact, or poor structural preservation that is visible on H&E but was not checked before submission. A pre-submission H&E or DAPI image reviewed by a pathologist or experienced histotechnologist catches most of these issues before samples ship.

References

1. Chen A, Liao S, Cheng M, et al. Spatiotemporal transcriptomic atlas of mouse organogenesis using DNA nanoball-patterned arrays. Cell. 2022;185(10):1777-1792.e21. doi:10.1016/j.cell.2022.04.003
2. You Y, Fu Y, Li L, et al. Systematic comparison of sequencing-based spatial transcriptomic methods. Nature Methods. 2024;21:1743-1754. doi:10.1038/s41592-024-02325-3
3. Zhao Y, Li Y, He Y, et al. Stereo-seq V2: spatial mapping of total RNA on FFPE sections with high resolution. Cell. 2025;188(23):6554-6571.e21. doi:10.1016/j.cell.2025.08.008
4. Grases D, Porta-Pardo E. A practical guide to spatial transcriptomics: lessons from over 1000 samples. Trends in Biotechnology. 2026;44(5):1230-1242. doi:10.1016/j.tibtech.2025.08.020
5. Zhang X, Zhang M, Xu Y, et al. ST-FFPE-mIF: integrating spatial transcriptomics and multiplex immunofluorescence in formalin-fixed paraffin-embedded tissues using Stereo-seq. Genome Biology. 2025;26:428. doi:10.1186/s13059-025-03900-3

The information provided in this article is for research use only and is not intended for use in diagnostic or therapeutic procedures. CD Genomics provides sequencing and bioinformatics services for research purposes. Researchers should consult the appropriate regulatory guidelines for their specific applications.