Figure 1: Overview of the single-nucleus RNA-seq workflow, from tissue to nuclear transcriptomic data.

Summary

Single-nucleus RNA sequencing (snRNA-seq) profiles gene expression from isolated nuclei rather than intact cells, making it a practical choice for frozen, archived, and difficult-to-dissociate tissue samples. This article explains what the nuclear transcriptome captures, how it differs from a whole-cell transcriptome, why intronic reads are central to snRNA-seq data, and which tissue contexts benefit most from a nuclei-first approach. Researchers evaluating single-cell methods for tissue studies will find clear, evidence-based guidance without spatial integration angles.

Key Takeaways

• snRNA-seq measures the RNA present inside the nucleus at the moment of isolation, capturing nascent transcripts, intronic reads, and transcripts undergoing processing that whole-cell methods may miss or distort.
• Nuclear transcriptomes differ from whole-cell transcriptomes in composition: they are enriched for intronic reads, exclude cytoplasmic RNA, and are not confounded by dissociation-induced stress artifacts.
• SnRNA-seq is especially useful for frozen tissues, large or fibrous tissues that resist intact-cell dissociation, and archived samples where whole-cell viability cannot be guaranteed.
• Cell-type recovery in snRNA-seq is not uniform across all types; certain cell populations (e.g., microglia, some epithelial subtypes) may be underrepresented relative to scRNA-seq.
• SnRNA-seq does not replace scRNA-seq for every project; the choice depends on sample state, the biology being studied, and what the research question actually requires.

What SnRNA-Seq Measures

Single-nucleus RNA sequencing starts with a simple but consequential shift in sample preparation: instead of isolating intact cells, the protocol isolates nuclei. This means the RNA that gets captured is the RNA that was inside the nucleus at the moment of tissue processing — not the full cellular transcriptome.

A nuclear transcriptome is fundamentally different from a whole-cell transcriptome. In a whole cell, the majority of captured RNA is mature cytoplasmic mRNA that has already been spliced, polyadenylated, and exported from the nucleus. In a nucleus, you get a different snapshot: nascent transcripts still being transcribed, pre-mRNA that has not yet undergone splicing, and a substantial fraction of intronic reads. This is not noise. Intronic reads in snRNA-seq data carry biological signal — they reflect transcriptional activity at the moment of capture and can reveal genes that are actively being transcribed even when mature mRNA levels are low.

Researchers new to snRNA-seq sometimes worry that intronic reads will degrade cell-type resolution. The evidence points the other way. Several studies have shown that counting intronic reads alongside exonic reads improves cell clustering and cell-type discrimination in snRNA-seq data, precisely because intronic signal captures genes in the act of transcription rather than only genes whose mature transcripts have accumulated in the cytoplasm.

The practical takeaway: snRNA-seq measures transcriptional activity rather than steady-state mRNA abundance. For many research questions — especially those concerned with gene regulation, dynamic transcriptional responses, or cell states in tissues where steady-state mRNA is sparse — this is a strength, not a limitation.

What is single-nucleus RNA-seq? It is a transcriptomic profiling method that sequences RNA from isolated single nuclei instead of intact single cells. It captures nascent and partially processed nuclear transcripts and is particularly suited for frozen, fibrotic, or archived tissue samples where whole-cell dissociation is unreliable or introduces artifacts.

Nuclear vs Whole-Cell Transcriptomes

Figure 2: Key differences between nuclear and whole-cell transcriptomes and their implications for data interpretation.

The differences between nuclear and whole-cell transcriptomes are not subtle. A 2020 systematic comparison of matched snRNA-seq and scRNA-seq from mouse kidney found that while both methods recovered broadly similar cell-type classifications, the gene expression profiles showed consistent, method-specific differences driven by transcript localization and processing state. Cytoplasmic transcripts — often the dominant signal in scRNA-seq — are simply absent from snRNA-seq data.

Three distinctions matter most for data interpretation.

First, gene detection sensitivity differs by gene class. Genes whose transcripts are rapidly exported from the nucleus after splicing tend to have lower counts in snRNA-seq. Genes with longer nuclear retention times, or genes that are actively transcribed at high rates, tend to be better represented. This means that when you compare snRNA-seq and scRNA-seq datasets directly, differences in per-gene counts may reflect biology (nuclear retention) rather than genuine expression differences.

Second, mitochondrial reads behave differently. In whole-cell scRNA-seq, a high mitochondrial read fraction often signals dying or stressed cells. In snRNA-seq, mitochondrial reads are lower overall because mitochondria are mostly cytoplasmic. This changes the QC metrics you use: the typical scRNA-seq filtering threshold for mitochondrial content does not translate directly to snRNA-seq data.

Third, ribosomal and non-coding RNA proportions shift. Nuclei contain ribosomal RNA in the nucleolus and various non-coding RNAs involved in splicing and processing. The transcriptional landscape in a nucleus has a different baseline composition than a whole cell, which affects normalization, clustering, and differential expression analysis if not accounted for.

These differences do not make one method better than the other in absolute terms. They make them different tools for different contexts. Understanding what each measures — and what each misses — is the foundation of choosing well.

Internal link: For researchers planning tissue-based studies, our snRNA sequencing services support nuclei isolation, library preparation, and bioinformatics analysis for a wide range of sample types.

The Role of Intronic Reads

Intronic reads are a defining feature of snRNA-seq data, and they are one of the most misunderstood aspects of the method. In scRNA-seq, intronic reads are often treated as contamination or noise and are filtered out during preprocessing. In snRNA-seq, they are an integral part of the signal.

The reason is straightforward: nuclei contain pre-mRNA. When a gene is actively transcribed, the primary transcript includes introns that have not yet been spliced out. Capturing these intronic reads means capturing transcriptional activity that may not yet be reflected in mature mRNA levels. This is especially informative for genes that are rapidly induced — their intronic signal can appear before cytoplasmic mRNA levels rise detectably.

A 2022 study examining artifactual gene expression signatures in brain tissue dissociates found that enzymatic tissue dissociation induced stress-response genes in microglia within minutes of processing. These artifacts appeared in whole-cell scRNA-seq data but were largely absent in matched snRNA-seq data. The nucleus, protected from the dissociation cocktail, preserved a transcriptional snapshot closer to the in vivo state.

For computational analysis, the standard practice is to include both exonic and intronic reads when building count matrices from snRNA-seq data. Tools like Cell Ranger (10x Genomics) support intronic read counting with the --include-introns flag, and several analysis workflows have been developed specifically to handle nuclear transcriptomic data. The Bioconductor OSCA workflow, for example, provides dedicated guidance on processing snRNA-seq data, including intronic read handling.

The practical implication for study design: if your research question involves transcriptional dynamics, stress responses, or gene regulatory programs that change rapidly, snRNA-seq's intronic reads may give you a more faithful readout than cytoplasmic mRNA alone.

Why Frozen Tissues Need Nuclei

Figure 3: Tissue types and sample states where single-nucleus RNA-seq provides practical advantages over whole-cell approaches.

One of the most common reasons researchers turn to snRNA-seq is sample logistics. Many tissue banks, clinical collections, and multi-center studies have extensive frozen tissue archives — but frozen tissues do not yield viable intact cells for scRNA-seq. Thawing and dissociating frozen tissue into single-cell suspensions typically produces dead cells, debris, and stress-induced artifacts that overwhelm biological signal.

Nuclei, in contrast, survive freezing well. The nuclear membrane provides structural protection, and simple mechanical homogenization in a detergent-based lysis buffer is often sufficient to release intact nuclei from flash-frozen tissue. This makes snRNA-seq the method of choice for frozen tissue cohorts that cannot be re-collected as fresh samples.

The advantage extends beyond frozen archives. Several tissue types are inherently difficult to dissociate into intact single cells, regardless of whether the sample is fresh or frozen. These include:

• Brain tissue, where extensive neuronal processes and myelin make whole-cell dissociation challenging, and where enzymatic digestion is known to induce artifactual gene expression in glial populations.
• Cardiac and skeletal muscle, where the structural syncytium resists single-cell suspension preparation.
• Fibrotic or tumor-associated stroma, where extracellular matrix stiffness impedes enzymatic digestion and yields poor cell viability.
• Adipose tissue, where lipid content and fragile adipocytes complicate standard dissociation protocols.

In all of these contexts, nuclei-based workflows bypass the dissociation bottleneck. The nucleus is isolated directly from the tissue, without requiring the cell body to remain intact through enzymatic or mechanical stress.

Protocols for nuclei isolation from frozen human tissues — including brain, kidney, and heart — have been well established and validated. A 2022 protocol from Current Protocols demonstrated efficient nuclei isolation from flash-frozen human heart tissue collected during surgery, and similar protocols exist for archived kidney biopsies and postmortem brain samples.

Internal link: If you are working with tissue biopsies and need to plan sample allocation, see our guide on Biopsy Material Allocation for Spatial Transcriptomics and snRNA-seq.

Cells SnRNA-Seq Finds Reliably

Not all cell types are equally recoverable by snRNA-seq. Understanding which cell types the method captures well — and which it underrepresents — is essential for study design and data interpretation.

Broadly, snRNA-seq recovers cell types with large, RNA-rich nuclei proportionally well. This includes neurons, cardiomyocytes, hepatocytes, and many stromal cell types. These are often the same cell types that are underrepresented or lost entirely during whole-cell dissociation, making snRNA-seq complementary rather than inferior.

However, certain cell populations are systematically underrepresented in snRNA-seq. In brain tissue, microglia — which have small, compact nuclei — are often recovered at lower proportions than expected from scRNA-seq or immunohistochemistry. A 2022 Nature Neuroscience study demonstrated that enzymatic dissociation introduced an aberrant ex vivo gene expression signature in microglia within minutes, while snRNA-seq largely avoided this artifact — but at the cost of reduced microglial recovery. The tradeoff is real: you avoid the stress signature but may capture fewer microglia overall.

Similarly, certain immune cells — particularly granulocytes and some lymphocyte subsets — can be underrepresented in snRNA-seq due to low nuclear RNA content or nuclear fragility. Epithelial cells in some tissues also show variable recovery depending on nuclear isolation conditions.

For researchers designing a study, the practical guidance is: if your biological question centers on a cell type known to have low nuclear RNA content or a small nucleus, validate its recovery in snRNA-seq pilot data before committing to a large-scale experiment. In many cases, running matched snRNA-seq and scRNA-seq on a subset of samples provides the most complete picture — snRNA-seq for cell types that dissociate poorly, scRNA-seq for cell types with low nuclear recovery.

The broader point is that snRNA-seq cell-type recovery is predictable when the tissue context is known. Researchers who understand their tissue's cellular composition can anticipate which populations will be well sampled and which may need complementary approaches.

What SnRNA-Seq Leaves Out

Like any method, snRNA-seq has limitations that should be weighed honestly during study design.

The most obvious gap is cytoplasmic RNA. Mature mRNA that has been exported from the nucleus is simply not captured. This means that for genes with rapid nuclear export and long cytoplasmic half-lives, snRNA-seq may underestimate expression levels relative to the true cellular transcriptome. Researchers studying RNA localization, translational regulation, or cytoplasmic RNA granules should be aware that snRNA-seq provides a nuclear-centric view.

A second limitation is splice isoform detection. Because nuclei contain pre-mRNA with unspliced introns, full-length isoform reconstruction is more challenging from snRNA-seq data than from scRNA-seq data, which primarily captures spliced cytoplasmic mRNA. Tools for isoform analysis generally assume mature transcripts, and their performance on nuclear data varies.

Third, cell-type recovery is not uniform, as discussed in the previous section. If your study requires comprehensive profiling of all cell types in a tissue — including those with small nuclei or low nuclear RNA content — snRNA-seq alone may not be sufficient.

Fourth, nuclei isolation introduces its own artifacts. Mechanical homogenization can shear nuclei, releasing nuclear contents and creating ambient RNA. Detergent-based lysis conditions that work for one tissue may not work for another. Nuclear membrane integrity after isolation affects data quality, and protocol optimization on a per-tissue basis is often necessary.

Finally, the analysis workflow differs from scRNA-seq. Standard scRNA-seq preprocessing pipelines that filter on mitochondrial content, gene count, and total molecular marker count need recalibration for nuclear data. Researchers should use analysis parameters and QC thresholds validated for snRNA-seq rather than importing scRNA-seq defaults.

None of these limitations should discourage researchers from using snRNA-seq — the method has proven its value across neuroscience, cardiology, oncology, and developmental biology. What these limitations mean is that snRNA-seq is a deliberate choice, not a default. It is the right tool when the sample demands it and the research question is compatible with what nuclear transcriptomics can deliver.

FAQ

What is single-nucleus RNA-seq and how is it different from scRNA-seq?

Single-nucleus RNA-seq (snRNA-seq) profiles gene expression from isolated single nuclei rather than intact single cells. The main difference is that snRNA-seq captures nuclear RNA — including nascent transcripts and intronic reads — while scRNA-seq captures primarily mature cytoplasmic mRNA from whole cells. This makes snRNA-seq better suited for frozen or difficult-to-dissociate tissues, but it also means the transcriptome it measures has a different composition than a whole-cell transcriptome.

When should I use snRNA-seq instead of scRNA-seq?

Choose snRNA-seq when your tissue cannot be dissociated into viable single cells without introducing stress artifacts, when you are working with frozen or archived samples, or when your cell type of interest is selectively lost during whole-cell dissociation (common for neurons, cardiomyocytes, and adipocytes). SnRNA-seq is also preferred when you want to capture transcriptional activity rather than steady-state mRNA levels, since intronic reads reflect active transcription.

Do intronic reads help or hurt cell-type classification in snRNA-seq?

Intronic reads generally improve cell-type classification in snRNA-seq. Because intronic reads capture actively transcribed genes, they provide additional signal that helps distinguish closely related cell types and cell-state transitions. Most snRNA-seq analysis workflows recommend including intronic reads in count matrices, and tools like Cell Ranger support intronic counting with the --include-introns option.

Can snRNA-seq detect all the same cell types as scRNA-seq?

No. SnRNA-seq recovers cell types with large, RNA-rich nuclei (e.g., neurons, cardiomyocytes, hepatocytes) proportionally well, but may underrepresent cell types with small nuclei or low nuclear RNA content (e.g., microglia, some immune subsets). Researchers should validate cell-type recovery in pilot experiments and consider matched scRNA-seq for populations with known low nuclear recovery.

Is snRNA-seq suitable for FFPE samples?

Most standard snRNA-seq protocols are optimized for fresh or flash-frozen tissue. FFPE samples present additional challenges because formalin fixation cross-links nucleic acids and degrades RNA. However, specialized protocols for FFPE nuclei isolation and snRNA-seq have been developed and are an active area of method development. Researchers interested in FFPE snRNA-seq should consult recent protocol literature and work with a service provider experienced in FFPE sample handling.

References

1. Grindberg RV, Yee-Greenbaum JL, McConnell MJ, et al. "RNA-sequencing from single nuclei." Proceedings of the National Academy of Sciences. 2013;110(49):19802–19807. DOI: 10.1073/pnas.1319700110
2. Denisenko E, Guo BB, Jones M, et al. "Systematic assessment of tissue dissociation and storage biases in single-cell and single-nucleus RNA-seq workflows." Genome Biology. 2020;21(1):130. DOI: 10.1186/s13059-020-02048-6
3. Marsh SE, Walker AJ, Kamath T, et al. "Dissection of artifactual and confounding glial signatures by single-cell sequencing of mouse and human brain." Nature Neuroscience. 2022;25(3):306–316. DOI: 10.1038/s41593-022-01022-8
4. Kim N, Kang H, Jo A, Yoo SA, Lee HO. "Perspectives on single-nucleus RNA sequencing in different cell types and tissues." Journal of Pathology and Translational Medicine. 2023;57(1):52–59. DOI: 10.4132/jptm.2022.12.19
5. Ding J, Adiconis X, Simmons SK, et al. "Systematic comparison of single-cell and single-nucleus RNA-sequencing methods." Nature Biotechnology. 2020;38(6):737–746. DOI: 10.1038/s41587-020-0465-8
6. Machado L, Relaix F, Mourikis P. "Stress relief: emerging methods to mitigate dissociation-induced artefacts." Trends in Cell Biology. 2021;31(11):888–897. DOI: 10.1016/j.tcb.2021.05.004
7. Slyper M, Porter CBM, Ashenberg O, et al. "A single-cell and single-nucleus RNA-Seq toolbox for fresh and frozen human tumors." Nature Medicine. 2020;26(5):792–802. DOI: 10.1038/s41591-020-0844-1