Preparing High-Viability Single-Cell Suspensions for scRNA-seq

Preparing a high-viability single-cell suspension for scRNA-seq is the step that decides how much usable data a sequencing run returns. Clustering, rare-population detection and differential expression all inherit whatever damage, bias or stress response the dissociation introduced, and no computational correction removes it completely. This protocol covers adherent cultures, organoids and solid tissue, with the quality checks that catch a bad suspension before it reaches the instrument rather than after sequencing.

What high viability means for scRNA-seq — and what it does not

Viability is a necessary release criterion, not a sufficient one. A suspension can read 95% viable and still give poor libraries, because cells that died earlier in the workflow have already leaked mRNA into the buffer. That free mRNA is captured alongside real cells as ambient or “soup” RNA, and it is present to some degree in every droplet dataset (Young & Behjati, GigaScience 2020). Four properties matter, and they are not interchangeable:

• Viability. Most droplet workflows are run with >80–90% viable cells. Confirm the current figure in your platform’s cell preparation guide rather than copying a number out of a paper.
• Cleanliness. Debris, free nuclei and membrane fragments raise ambient RNA and blur cluster boundaries.
• Singularity. Clumps cause doublets and channel blockages. Aim for mostly single cells with occasional doublets — a suspension with literally no doublets is usually over-digested.
• Fidelity. The transcriptome you sequence should resemble the cells in the dish or the tissue, not their response to being removed from it. That is the artefact discussed further down, and it is the one most often missed.

Materials and decisions before you start

Decide these four things before touching the cells. Changing them mid-experiment is the most common reason two replicates do not agree.

1. Dissociation reagent. For adherent lines and organoids, a trypsin-like recombinant enzyme (TrypLE™-type or an animal-origin-free recombinant trypsin-like enzyme such as CellTrypase) or Accutase®. For solid tissue, an enzyme matched to the matrix — see our tissue dissociation enzyme selection guide. Browse the full enzymes collection for options.
2. Temperature. Warm digestion (37 °C) is faster and induces a conserved stress-gene response; cold digestion with a psychrophilic (cold-active) protease at about 6 °C largely suppresses it (Adam et al., Development 2017; O’Flanagan et al., Genome Biology 2019).
3. DNase I. Include it from the start of the enzymatic step wherever cells will lyse — typically 10–100 µg/mL. Free genomic DNA is the usual cause of stubborn clumping.
4. Endpoint. Fresh cells, cryopreserved cells, fixed cells or nuclei behave differently. Cryopreservation of a dissociated suspension can selectively lose epithelial cell types, and methanol fixation preserves composition but increases ambient RNA (Denisenko et al., Genome Biology 2020).

Also prepare: Ca²⁺/Mg²⁺-free PBS; wash buffer (commonly PBS with 0.04% BSA — check your platform’s guide); wide-bore or cut pipette tips; low-binding tubes; 37–40 µm cell strainers, pre-wetted; a haemocytometer or automated counter with a viability dye; and 10 µM Y-27632 if single cells will be replated (Watanabe et al., Nature Biotechnology 2007).

Protocol A — adherent cultures and cell lines

1. Plan the harvest. Work with cells at roughly 70–80% confluence in log phase. Over-confluent monolayers need longer digestion and give lower viability.
2. Pre-warm the dissociation reagent and medium to room temperature or 37 °C. Cold reagent on a cold monolayer is the most common cause of a 20-minute detachment that should have taken 3 minutes.
3. Wash. Aspirate spent medium and rinse the monolayer once with Ca²⁺/Mg²⁺-free PBS to remove serum, which inhibits the enzyme.
4. Add the enzyme to cover the cell layer — approximately 1 mL per 25 cm² of growth area.
5. Incubate and watch. Check on the microscope every 1–2 minutes. Stop as soon as the cells are rounded and released; a gentle tap on the flask should be enough. Do not run the incubation to a fixed time taken from another cell line.
6. Quench. Add 3–5 volumes of medium. Serum-free workflows can quench by dilution in buffer or medium alone where the enzyme allows it — recombinant trypsin-like enzymes of fungal origin do not require a trypsin inhibitor.
7. Disperse gently. Pipette 5–10 times with a wide-bore tip. Aggressive pipetting shears cells and is a major, avoidable source of ambient RNA.
8. Pellet at 300 × g for 5 minutes at 4 °C. Resuspend in cold wash buffer. Repeat once if enzyme carryover matters for your downstream step.
9. Strain through a pre-wetted 37–40 µm strainer.
10. Count twice and record viability. Keep the suspension on ice and load as soon as the platform allows — viability falls with time on the bench.

Protocol B — solid tissue, organoids and primary material

1. Keep it cold until digestion. Transport tissue in cold basal medium and process as quickly as possible. Elevated temperature alone is enough to induce an artefactual activation signature in human brain tissue (Marsh et al., Nature Neuroscience 2022).
2. Mince to fragments of about 1 mm³ with crossed scalpels in a small volume of cold buffer. Mechanical work done well here shortens the enzymatic step.
3. Organoids only: remove the extracellular matrix first with two washes in ice-cold basal medium or a cell-recovery solution, 300 × g, 5 minutes, 4 °C. Matrix carryover blocks strainers and traps cells. See the dedicated organoid dissociation protocol.
4. Digest. Add the chosen enzyme mix with DNase I. Either digest warm at 37 °C with agitation, accepting the stress signature, or digest cold with a cold-active protease at about 6 °C for 10–30 minutes, which is slower but preserves the in vivo transcriptional state far better.
5. Check every 3–5 minutes under the microscope. Stop when the field is mostly single cells with occasional doublets.
6. Quench and filter. Dilute in cold buffer, pass through 70 µm, then 40 µm strainers.
7. Clean up. Lyse red blood cells if present; remove debris on a density gradient; remove dead cells magnetically or by annexin-based depletion if viability is below target.
8. Final wash at 300–400 × g, 5 minutes, 4 °C, and resuspend in wash buffer at the concentration your platform specifies.

Removing debris, dead cells and doublets

Three cleanup decisions repay the extra 20 minutes:

• Dead-cell depletion when viability is between roughly 60% and 85%. Below that, the suspension is usually not rescuable and the dissociation should be redesigned.
• Debris removal by density gradient for fibrous or fatty tissue. Debris is invisible in a viability count but very visible in the ambient RNA profile.
• Re-straining immediately before loading. Suspensions re-aggregate on ice within 30 minutes, especially when free DNA is present.

Doublets that survive into the library can be flagged computationally afterwards — for example with DoubletFinder — and ambient RNA can be estimated and subtracted with SoupX. Neither is a substitute for a clean suspension; both work better on one.

Quality checks before you load

1. Two independent counts within 10% of each other. If they disagree, the suspension is not homogeneous — mix gently and recount.
2. Viability at or above your platform’s specification, measured with a dual stain rather than trypan blue alone where possible.
3. Visual check of 5 µL on a slide: single cells, intact membranes, no visible clumps or fibres.
4. Concentration adjusted to the platform’s target range. Loading above range raises the doublet rate proportionally.
5. Time on ice recorded. Treat it as a variable to keep constant across samples in a batch, not as an incidental detail.

The dissociation artefact: how to control it and how to detect it

Warm enzymatic dissociation switches on an immediate-early and heat-shock transcriptional programme. Working across 155,165 cells from patient tumours, xenografts and cell lines, O’Flanagan and colleagues defined a core set of 512 stress and heat-shock genes — including FOS and JUN — induced by collagenase digestion at 37 °C and largely absent when the same tissue was dissociated with a cold-active protease at 6 °C. They also note that expression differences in immune-recognition components such as MHC class I may be especially confounded by this effect.

The same phenomenon has been described independently in muscle stem cells (van den Brink et al., Nature Methods 2017), in developing kidney (Adam et al., 2017), and in brain, where enzymatic dissociation produces a prominent artefactual microglial activation signature that is present in a large fraction of the published literature (Marsh et al., 2022). Practical consequences:

• Score the artefact, do not assume it is absent. Calculate a module score for the 512-gene set in your own data before interpreting any stress, activation or hypoxia cluster as biology.
• Keep dissociation constant across conditions. A treated and a control sample dissociated on different days differ by their dissociation as well as by treatment.
• Consider single-nucleus RNA-seq for fragile or hard-to-release cell types. In adult kidney, snRNA-seq recovered podocytes at 2.4% versus 0.12% in published scRNA-seq data and showed no dissociation stress genes, at the cost of underrepresenting T, B and NK cells (Wu et al., JASN 2019; Denisenko et al., 2020).

Troubleshooting a failed suspension

Where CellTrypase fits — and where the data stops

CellTrypase (Kerry, formerly c-LEcta; BioHippo cat. BHZ16500003 / art. 22103-1X-100) is a recombinant, animal-origin-free trypsin-like serine protease. The gene comes from Fusarium oxysporum and is expressed in a Bacillus sp. host; the enzyme is about 22 kDa and cleaves after lysine and arginine. It is released at ≥95% purity by HPLC, supplied sterile-filtered in PBS with 1.1 mM EDTA at pH 7.1–7.6 and 270–320 mOsm/kg, with endotoxin ≤1 EU/mL at 1× and ≤10 EU/mL at 10×, mycoplasma negative, and sterility passing Ph. Eur. 2.6.1 / USP <71>. It is quenched by dilution in buffer or medium — no trypsin inhibitor is added to the suspension — and the manufacturer confirms stability of at least 21 months at 2–8 °C, with shipment at ambient temperature.

The manufacturer’s performance data covers four adherent production lines, measured against an unnamed current industry-standard trypsin-like enzyme as control:

Source: c-LEcta / Kerry, CellTrypase Product Information Sheet v4.0 (18 Aug 2026). MDCK is shown as published: 23:32 is a long release time for that line, and it is reported here unaltered so you can judge it against your own passaging times.

What that table does not tell you is anything about single-cell sequencing. There is no published data — from the manufacturer or from us — on how CellTrypase affects viability, ambient RNA, recovery or the dissociation stress signature in an scRNA-seq workflow, and none for primary cells, tumours, organoids or neurons. Manufacturer performance data exists for CHO-K1, HEK 293, MDCK and Vero, plus 2D and 3D human iPSC data in the Kerry technical documents. Everything else is an intended application without performance data behind it. We would rather say that than imply a number we do not have.

If you want to know how it behaves in your model, the honest route is to test it: a free 100 mL sample is available for qualification, R&D or GMP grade. Run it beside your current reagent on the same cells, on the same day, and compare release time, viability, recovery and — if the endpoint is sequencing — the stress-gene module score. For dosing, 1 mL per 25 cm² and the 1× format suit most applications; if you substitute it into an existing protocol at the same volume, verify detachment time and viability in your own cell model before switching.

Related reading: CellTrypase vs TrypLE™ — a reagent comparison and the CellTrypase cell dissociation protocol. For cell models and primary material, see the cell lines collection.

TrypLE™ is a trademark of Thermo Fisher Scientific. Accutase® is a registered trademark of its respective owner. Trademarks are used here for identification and comparison only; no endorsement or affiliation is implied. CellTrypase R&D grade is for research use only (RUO) — not for diagnostic or therapeutic use.

References

1. van den Brink SC, et al. Single-cell sequencing reveals dissociation-induced gene expression in tissue subpopulations. Nat Methods. 2017;14(10):935–936. doi:10.1038/nmeth.4437
2. Adam M, Potter AS, Potter SS. Psychrophilic proteases dramatically reduce single-cell RNA-seq artifacts: a molecular atlas of kidney development. Development. 2017;144(19):3625–3632. doi:10.1242/dev.151142
3. O’Flanagan CH, et al. Dissociation of solid tumor tissues with cold active protease for single-cell RNA-seq minimizes conserved collagenase-associated stress responses. Genome Biol. 2019;20(1):210. doi:10.1186/s13059-019-1830-0
4. Denisenko E, et al. Systematic assessment of tissue dissociation and storage biases in single-cell and single-nucleus RNA-seq workflows. Genome Biol. 2020;21(1):130. doi:10.1186/s13059-020-02048-6
5. Marsh SE, et al. Dissection of artifactual and confounding glial signatures by single-cell sequencing of mouse and human brain. Nat Neurosci. 2022;25(3):306–316. doi:10.1038/s41593-022-01022-8
6. Wu H, et al. Advantages of single-nucleus over single-cell RNA sequencing of adult kidney. J Am Soc Nephrol. 2019;30(1):23–32. doi:10.1681/ASN.2018090912
7. Young MD, Behjati S. SoupX removes ambient RNA contamination from droplet-based single-cell RNA sequencing data. GigaScience. 2020;9(12):giaa151. doi:10.1093/gigascience/giaa151
8. McGinnis CS, Murrow LM, Gartner ZJ. DoubletFinder: doublet detection in single-cell RNA sequencing data using artificial nearest neighbors. Cell Syst. 2019;8(4):329–337.e4. doi:10.1016/j.cels.2019.03.003
9. Watanabe K, et al. A ROCK inhibitor permits survival of dissociated human embryonic stem cells. Nat Biotechnol. 2007;25(6):681–686. doi:10.1038/nbt1310
10. Reichard A, Asosingh K. Best practices for preparing a single cell suspension from solid tissues for flow cytometry. Cytometry A. 2019;95(2):219–226. doi:10.1002/cyto.a.23690
11. CellTrypase Product Information Sheet v4.0, valid as of 18 Aug 2026. c-LEcta GmbH (Kerry Biotechnology Centre).