When to Choose Stereo-seq for Subcellular Spatial Transcriptomics

Inquiry

Figure 1. Conceptual overview of the Stereo-seq decision framework — matching project goals to platform capabilities across resolution, tissue coverage, and sample type.

Stereo-seq offers a combination no other spatial transcriptomics platform currently matches: subcellular resolution paired with a centimeter-scale capture area. But that combination does not make it the right choice for every project. This article walks through the specific scenarios where Stereo-seq outperforms alternatives — and the situations where another platform, or a hybrid approach, may serve you better.

Key Takeaways

• Stereo-seq's unique strength is resolution + field of view. With 500 nm center-to-center DNB spacing and capture areas up to 13.2 × 13.2 cm, it is the only platform that covers whole organs at subcellular resolution.
• It excels when you need unbiased, whole-transcriptome discovery. Unlike imaging-based platforms (Xenium, CosMx) with fixed gene panels, Stereo-seq captures poly(A) RNA without pre-selecting targets.
• Tissue type matters for diffusion control. Stereo-seq performs best in dense tissues (embryo, eye, liver); lateral diffusion is more pronounced in brain and other lipid-rich tissues.
• FFPE compatibility has improved but is still maturing. The Stereo-seq V2 random-primer workflow now supports FFPE, but the fresh-frozen protocol is more established with broader validation.
• Host-microbe co-detection is a distinctive strength that no other commercial subcellular-resolution platform currently provides.

What Stereo-seq Brings to the Table

Stereo-seq (Spatial Enhanced Resolution Omics-sequencing) builds on BGI's DNA nanoball (DNB) technology. Random-barcoded oligonucleotides are amplified into ~220 nm DNBs and patterned at 500 nm center-to-center spacing on a silicon chip. Before tissue is applied, the spatial barcode — called a Coordinate Identity (CID) — is decoded by 25 cycles of sequencing, so every capture spot carries a known physical address.

When a tissue section is placed on the chip and permeabilized, mRNA is captured in situ. After reverse transcription and sequencing on an MGI platform, each transcript is mapped back to its tissue coordinate. The result is a spatial gene expression matrix with subcellular resolution across a field of view that can reach 13.2 × 13.2 cm — roughly 500 times larger than a standard Visium capture area.

How Stereo-seq compares at a glance:

This comparison is intentionally brief. For a broader cross-platform discussion that also covers imaging-based methods and how to choose between sequencing and imaging approaches, see the spatial transcriptomics platform comparison guide.

Five Scenarios Where Stereo-seq Excels

Stereo-seq is not the default answer for every spatial project. But in five specific scenarios, it is the strongest candidate — and often the only platform that delivers what the study design requires.

Whole-Organ or Large-Area Mapping

If your study requires capturing an entire mouse embryo sagittal section, a whole human tissue slice several centimeters across, or an organ-scale developmental series, Stereo-seq is the only subcellular-resolution platform with a large enough capture area. The flagship Cell (2022) study demonstrated this by profiling 53 sagittal sections across eight developmental timepoints, generating the Mouse Organogenesis Spatiotemporal Transcriptomic Atlas (MOSTA). No other platform can cover 13 cm of continuous tissue at subcellular resolution in a single run. This capability matters practically: without it, large structures must be reconstructed from multiple smaller capture areas, introducing stitching artifacts and registration uncertainty that can obscure fine spatial boundaries between adjacent anatomical regions.

Discovery Without a Predefined Gene List

Imaging-based platforms require you to choose a gene panel before the experiment. If your project is exploratory — you want to discover which genes and pathways matter, not confirm a candidate list — Stereo-seq's unbiased poly(A) capture gives you the whole transcriptome. In the 2025 Nature Communications benchmarking study across human tumors, Stereo-seq detected the broadest gene coverage and showed the highest correlation with matched scRNA-seq data (R = 0.85), making it the strongest sST platform for discovery-oriented projects.

Host-Microbe Co-detection

No other commercial subcellular-resolution spatial platform can simultaneously capture host and microbial transcripts on the same tissue section. Stereo-seq's unbiased poly(A) capture detects both human and non-human RNA without requiring species-specific probes. This makes it uniquely suited for infectious disease research, microbiome-tissue interface studies, and host-pathogen interaction mapping — demonstrated most recently in the Cell (2025) Stereo-seq V2 study using Mycobacterium tuberculosis-infected tissue.

Non-Model Species and Cross-Species Studies

Because Stereo-seq relies on poly(A) capture rather than species-specific probe panels, it works across essentially any species with polyadenylated RNA. Researchers have applied it to mouse, human, zebrafish, Drosophila, Arabidopsis, axolotl, macaque, and rice — often without needing to develop new reagents. Imaging-based platforms, by contrast, typically require probe panels designed against specific genomes, limiting their use in non-model organisms. For plant researchers, this is a decisive factor: Stereo-seq has produced the first spatially resolved single-cell transcriptomic maps of Arabidopsis and maize leaves, revealing C3/C4 photosynthetic gene gradients and epidermal cell subtype organization that panel-based platforms could not have captured without prior knowledge of which genes to target.

Coding and Non-Coding RNA in a Single Readout

Stereo-seq's poly(A) capture recovers mRNA, lncRNA, and other polyadenylated non-coding transcripts in the same experiment. With the Stereo-seq V2 random-primer workflow for FFPE, this extends to total RNA — including non-polyadenylated species such as snoRNAs and circular RNAs. For projects where regulatory RNA biology is central to the question — enhancer RNA dynamics, imprinted lncRNA localization, or splicing variant spatial patterns — this breadth of RNA capture can reveal layers of spatial regulation that panel-based methods miss. The V2 workflow also enables detection of tumor-specific alternative splicing events from FFPE sections, linking spatial context to post-transcriptional regulation in archival cancer specimens.

Figure 2. Five research scenarios where Stereo-seq provides capabilities that competing subcellular-resolution platforms cannot match.

Where the Limits Show Up

No platform is universal. Knowing where Stereo-seq underperforms is as important as knowing where it excels — it prevents wasted samples and uninterpretable data.

Tissue-type-dependent diffusion. In the 2024 Nature Methods benchmark of 11 sequencing-based methods, Stereo-seq showed excellent diffusion control in mouse embryo and eye but markedly stronger lateral diffusion in mouse brain (olfactory bulb, hippocampus). The 2025 multi-tumor benchmark confirmed this pattern: diffusion is tissue-architecture-dependent. Dense, cell-rich tissues fare well; lipid-rich or loosely structured tissues can produce transcript spreading that blurs fine spatial boundaries. If your tissue type has not been tested with Stereo-seq, a small pilot section is a prudent first step.

MGI sequencer requirement. Stereo-seq libraries are sequenced on MGI's DNBSEQ platform (typically T7). If your lab or service provider uses Illumina sequencers exclusively, Stereo-seq adds a platform dependency. This matters for project planning: the sequencing step cannot be redirected to a different instrument.

FFPE maturity. Stereo-seq's FFPE workflow (V2, random-primer based) was published in Cell in mid-2025 and represents a genuine advance — including total RNA capture and host-microbe co-detection from archived blocks. But the fresh-frozen protocol has several years more validation, more published benchmarks, and a larger user community. For FFPE projects, Visium HD and Xenium both have longer track records in clinical-translational settings.

Cell segmentation still evolving. Stereo-seq does not natively image cell boundaries — it relies on DAPI nuclear staining plus computational expansion (e.g., 5 μm nuclear dilation) to approximate cell areas. In the 2025 benchmark, this approach produced more false-positive segmentations from non-cellular structures compared to the multi-channel membrane staining used by Xenium and CosMx. If your analysis depends on precise single-cell boundaries (cell-cell communication, ligand-receptor mapping), an imaging-based platform may be more suitable, or you may plan to integrate Stereo-seq with a matched scRNA-seq reference for deconvolution and spatial mapping. The newer CellBin algorithm and nucleus-segmentation-based approaches have improved Stereo-seq single-cell resolution, but they remain computationally dependent in a way that direct membrane imaging is not.

Fresh Frozen, FFPE, or Both

Sample type is often the first constraint — and it directly affects what Stereo-seq can deliver.

A practical rule of thumb: if you are collecting samples prospectively for a discovery project, fresh frozen gives you the most validated Stereo-seq workflow. If you have archived FFPE blocks and your scientific question centers on total RNA, microbial co-detection, or retrospective cohorts, the Stereo-seq V2 FFPE workflow is now viable — but budget time for data quality review and consider running a test section before committing the full cohort.

For projects that depend on FFPE sample quality assessment before committing to spatial transcriptomics, the FFPE spatial transcriptomics service page covers additional feasibility considerations.

Asking the Right Questions Before You Start

Stereo-seq projects succeed or fail on the quality of the upfront decisions. Work through these questions before tissue touches a chip:

• What is the spatial scale of the biology you need to capture? If you need to see subcellular RNA localization (e.g., transcripts at the nuclear membrane vs. cytoplasm), Stereo-seq's 500 nm DNB spacing can resolve it. If you need single-cell precision for ligand-receptor analysis, an imaging-based platform or a scRNA-seq integration strategy may be more reliable.
• How large is the tissue region of interest? For regions under ~1 cm², multiple platforms are viable. For regions above 1 cm² in a single continuous section, Stereo-seq is the only subcellular-resolution option.
• Do you need to detect non-host RNA or non-coding RNA? If yes, Stereo-seq provides capture breadth that probe-based methods cannot replicate.
• Is your tissue dense and cell-rich, or loose and lipid-rich? Dense tissues (liver, embryo, tumor cores) typically produce cleaner Stereo-seq data. Brain, adipose, and fibrotic tissues benefit from a pilot test.
• What is your sample format — fresh frozen or FFPE? Fresh frozen is the lower-risk choice for Stereo-seq today. FFPE is feasible with the V2 workflow but has fewer published validation datasets.
• Do you have access to an MGI sequencer? If not, confirm that your service provider handles MGI-based sequencing as part of the Stereo-seq workflow.

For researchers ready to evaluate whether Stereo-seq fits a specific tissue and research question, CD Genomics Spatial Transcriptomics Services include Stereo-seq as one of multiple platform options, with sample feasibility review and study design support available before project commitment.

Figure 3. Pre-project decision checklist for researchers evaluating whether Stereo-seq fits their spatial transcriptomics study.

FAQ

Q: When is Stereo-seq the right choice over Visium HD?

Stereo-seq provides finer spatial resolution (500 nm vs. 2 μm bins) and a much larger maximum capture area (up to 13.2 cm vs. 6.5 mm per capture area). Visium HD offers better diffusion control in challenging tissues such as brain, does not require an MGI sequencer, and has a more mature FFPE workflow with longer track record in clinical-translational settings. If your project needs whole-transcriptome coverage and the tissue is dense and fresh frozen, Stereo-seq is often the stronger choice. If your project uses FFPE samples and single-cell resolution matters more than capture area, Visium HD may be more practical.

Q: Can Stereo-seq detect non-coding RNA?

Yes. The standard fresh-frozen Stereo-seq workflow uses poly(A) capture, which recovers mRNA and polyadenylated long non-coding RNAs. The newer Stereo-seq V2 FFPE workflow uses random primers and captures total RNA — including non-polyadenylated transcripts such as snoRNAs and certain regulatory RNAs. This makes V2 particularly useful for projects where non-coding RNA biology is central to the research question.

Q: How does Stereo-seq handle host-microbe co-detection in practice?

Because Stereo-seq captures RNA via poly(A) tails rather than species-specific probe hybridization, it detects polyadenylated transcripts from both host and microbial cells on the same tissue section without requiring separate probe panels. This was demonstrated in the 2025 Cell study on M. tuberculosis-infected tissue, where both human and bacterial transcripts were spatially mapped from a single FFPE section. The key practical requirement is sufficient sequencing depth to detect the typically lower-abundance microbial transcripts alongside the host transcriptome.

Q: What is the minimum tissue size for Stereo-seq to work well?

There is no strict minimum — Stereo-seq chips are available in sizes down to 0.5 × 0.5 cm. However, for very small regions of interest (e.g., a single glomerulus or a small tumor biopsy), the key question is whether subcellular resolution and whole-transcriptome coverage justify the platform choice. For tiny, well-defined structures where you already know which genes matter, an imaging-based platform with a targeted panel (Xenium, CosMx) may give you cleaner single-cell data with less diffusion-related ambiguity.

Q: What if my tissue type has not been tested with Stereo-seq?

Run a small pilot section first. Stereo-seq's performance is tissue-type-dependent — particularly regarding lateral diffusion and RNA capture efficiency. A pilot section processed with the same protocol you plan to use (fresh frozen or FFPE) will tell you whether the spatial signal is clean enough for your biological question before you commit an entire cohort or a precious large section.

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. Ren P, Zhang R, Luo C, et al. Systematic benchmarking of high-throughput subcellular spatial transcriptomics platforms across human tumors. Nature Communications. 2025;16:9232. doi:10.1038/s41467-025-64292-3
4. 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
5. 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

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.