10x Visium FFPE Spatial Transcriptomics Service for Archived Tissue
CD Genomics provides probe-based FFPE spatial transcriptomics services that map whole-transcriptome gene expression directly from archived formalin-fixed paraffin-embedded tissue sections. Unlike poly-A capture methods that require intact mRNA, our approach uses paired oligonucleotide probe hybridization and ligation — designed for the fragmented RNA typical of FFPE specimens — to profile over 18,000 protein-coding genes while preserving spatial coordinates across the tissue. From decades-old clinical blocks to freshly embedded research specimens, we deliver spatially resolved gene expression data with histological context, supporting retrospective cohort studies, biomarker discovery, and translational research at tissue-level and near-single-cell resolution.
Why CD Genomics for FFPE spatial transcriptomics:
- Probe-based chemistry compatible with degraded FFPE RNA — no DV200 cutoff required for project initiation; samples with DV200 values well below 50% have been successfully profiled
- Whole-transcriptome coverage of >18,000 genes per tissue section with spatial coordinates preserved — from discovery through validation in a single experiment
- Multi-resolution options spanning tissue-level (55 µm spots) to near-single-cell (8 µm bins) analysis, matched to your biological question
- Validated across FFPE blocks stored for over a decade, multiple carcinoma types, and diverse tissue origins — including specimens from retrospective clinical archives
- Integration-ready data compatible with single-nucleus RNA-seq, spatial transcriptomics (fresh frozen), and spatial proteomics for multi-modal tissue atlasing
- End-to-end workflow from tissue section QC through sequencing and publication-grade bioinformatics, managed by a single team
Services are provided for research use only.
Technology Overview — Probe-Based Spatial Transcriptomics for FFPE
Why FFPE Tissue Requires a Different Approach
Standard spatial transcriptomics workflows that rely on poly-A capture of mRNA tails fail on FFPE samples because formalin fixation crosslinks RNA to proteins and introduces strand breaks — producing RNA fragments typically 200–500 nucleotides long with chemically modified bases. Paraffin embedding and long-term storage further degrade RNA. The result: intact poly-A tails are rare, and the full-length cDNA synthesis step central to poly-A-based methods produces little usable signal from FFPE material.
Probe-based spatial transcriptomics bypasses this limitation. Instead of capturing mRNA through its poly-A tail, the method uses pairs of short oligonucleotide probes (~25 nucleotides each) that hybridize to adjacent sequences within each target gene transcript. Because each probe pair requires only ~50 nucleotides of contiguous accessible RNA, the method is effective even on heavily fragmented FFPE specimens — including blocks archived for 10–20 years.
Probe-Based vs. Poly-A Capture: Why Probe Chemistry Matters for FFPE
| Factor | Probe-Based (FFPE-Optimized) | Poly-A Capture (Fresh Frozen) |
|---|---|---|
| RNA requirement | ~50 nt accessible fragments per probe pair | Full-length mRNA with intact poly-A tail |
| FFPE compatibility | Yes — designed for fragmented, crosslinked RNA | No — FFPE RNA is too degraded for reliable poly-A capture |
| Gene coverage | >18,000 genes (whole transcriptome probe panel) | >18,000 genes (theoretical; depends on RNA integrity) |
| Block age tolerance | Successfully demonstrated on blocks archived 10+ years | Not applicable to archival material |
| SNP detection in probe regions | Not applicable — reads align to probe reference | SNPs detectable in captured mRNA sequence |
| Coverage uniformity | Probe design ensures even coverage across targeted genes | Coverage varies with transcript abundance and 3' bias |
How Probe-Based Spatial Transcriptomics Works
The FFPE spatial transcriptomics workflow proceeds through five linked molecular events on a single tissue section:
- Probe hybridization — A panel of paired DNA oligonucleotide probes (targeting >18,000 genes across the human transcriptome) is applied to a deparaffinized, decrosslinked FFPE tissue section. Probes hybridize overnight to their complementary target sequences in the fragmented mRNA pool. For human samples, three probe pairs per gene improve capture sensitivity.
- Probe ligation — After washing away unbound probes, hybridized probe pairs that sit adjacent to one another on the same mRNA molecule are enzymatically ligated into a single DNA fragment. Ligation requires both probe halves to independently hybridize to the same transcript — providing inherent specificity. Unpaired probes cannot ligate and are removed.
- Spatial capture — The tissue section with ligated probe products is aligned to a slide carrying an array of spatially barcoded capture oligonucleotides. Upon release from the tissue, the ligated probes hybridize to the spatial barcodes via a poly-A sequence on the right-hand probe — transferring each transcript's identity and spatial position to a known coordinate on the array.
- Library preparation — Spatially barcoded probe products are extended, eluted, PCR-amplified, and converted into sequencing libraries. The resulting libraries contain the spatial barcode, a molecular tag for deduplication, and the probe-target sequence.
- Sequencing and data processing — Libraries are sequenced on Illumina platforms. Computational pipelines map reads to a probe reference, count probe ligation events per spatial feature, and generate a feature-barcode expression matrix linked to tissue coordinates.
What the Data Reveals
| Information Layer | Biological Insight |
|---|---|
| Spatial gene expression maps | Which genes are expressed in each tissue domain |
| Unbiased transcriptome coverage | Discovery of spatially patterned genes without panel preselection |
| Histology-matched expression | Direct overlay of gene expression onto H&E morphology |
| Cell-type deconvolution | Estimation of cell-type composition per spatial feature using reference single-cell or single-nucleus data |
| Spatial neighborhoods | Identification of recurrent cellular communities and their gene expression programs |
| Ligand-receptor spatial coupling | Inference of cell-cell communication from co-localized expression of ligand-receptor pairs |
Standard Resolution vs. High-Definition Spatial Transcriptomics
| Factor | Standard Resolution | High-Definition (HD) |
|---|---|---|
| Spatial resolution | 55 µm spots (multicellular, ~1–20 cells) | 2 µm squares, analyzed at 8 µm bins (near-single-cell) |
| Discovery vs. resolution | Better for tissue-domain-level discovery and cohort studies | Better for cell-level spatial mapping within complex tissues |
| Sequencing cost | Lower (~25K reads/spot × 5,000 spots) | Higher (~275M reads per capture area) |
| Data analysis complexity | Standard spatial pipelines | Requires additional binning and higher-resolution analysis |
| Best for | Large-cohort screening, tissue architecture mapping | Rare cell detection, fine-grained spatial organization, tumor-immune interface mapping |
Service Workflow
Our FFPE spatial transcriptomics workflow integrates tissue QC, probe chemistry, spatial capture, sequencing, and bioinformatics in a single coordinated pipeline.
- Study Design and Tissue QC
Your project begins with a consultative planning phase where we align tissue type, specimen age, resolution option (standard or HD), sample number, and experimental design with your biological questions. FFPE blocks or pre-cut sections are assessed for morphology and RNA quality. QC Checkpoint: Sections with extensive necrosis (>30%), folding, or detachment are flagged and discussed before proceeding. DV200 is measured for informational purposes — low values do not automatically disqualify samples.
- Deparaffinization, Decrosslinking, and Probe Hybridization
FFPE sections are deparaffinized, rehydrated, and subjected to controlled decrosslinking to expose target RNA. The whole-transcriptome probe panel — covering >18,000 human genes — is hybridized to the section overnight. QC Checkpoint: Positive and negative control probe signals are monitored to assess hybridization uniformity across the tissue area.
- Probe Ligation and Spatial Capture
Unbound probes are washed away, and adjacent probe pairs hybridized to the same mRNA are enzymatically ligated. The tissue section is then aligned with a spatially barcoded capture slide. Ligated probe products are released and captured at their tissue coordinates via hybridization of the poly-A tail to spatially arrayed capture oligos. QC Checkpoint: Tissue imaging post-alignment confirms registration accuracy between the H&E image and the spatial barcode grid.
- Library Preparation and Sequencing
Captured probes are extended, eluted, PCR-amplified, and converted into Illumina-compatible sequencing libraries. Libraries undergo quality assessment before sequencing at the depth matched to your chosen resolution. QC Checkpoint: Library trace, Q30 scores, and sequencing saturation are monitored; under-sequenced samples are flagged for additional sequencing if needed.
- Bioinformatics Analysis and Data Delivery
Sequencing data are processed through a spatial transcriptomics pipeline: read alignment to the probe reference, spatial barcode demultiplexing, feature-barcode matrix generation, spatial clustering, differential expression analysis between tissue regions, cell-type deconvolution, and generation of publication-ready spatial expression maps. All analysis parameters are logged. QC Checkpoint: Final data review against project specifications before delivery.
Sample Requirements
| Requirement | Specification |
|---|---|
| Sample type | FFPE tissue blocks or pre-cut FFPE sections on adhesive slides |
| Species | Human, mouse (additional species evaluated case by case) |
| Section thickness | 5 µm (standard); 5–10 µm acceptable |
| Slide type | SuperFrost Plus or equivalent charged/adhesive slides |
| Tissue area | Standard capture area: 6.5 × 6.5 mm per section. Tissues larger than the capture area can be accommodated by selecting representative sub-regions |
| Minimum sections | 3–4 consecutive sections per sample recommended (1 for H&E/QC, 1–2 for spatial transcriptomics, 1 backup) |
| Block age | Blocks up to 10+ years old have been successfully processed. Older specimens are evaluated case by case |
| RNA quality | DV200 is recorded for QC documentation. Low DV200 is not an automatic exclusion criterion; samples with DV200 values below 30% have yielded interpretable spatial transcriptomic data in published studies |
| Morphology requirements | Sections should be intact without folding, tearing, or detachment. Necrotic regions and RBC-contaminated areas should each be below approximately 20% of the analysis area |
| Shipping | FFPE blocks or slides shipped at ambient temperature. Pre-cut sections should be shipped within 1–2 weeks of sectioning for optimal performance |
Tissue types successfully processed include breast carcinoma, non-small cell lung cancer, colorectal adenocarcinoma, diffuse large B-cell lymphoma, cervical tissue, kidney biopsies, and skin. For tissue types not previously validated, contact our team to discuss feasibility and protocol optimization.
For detailed sample preparation and shipping instructions, contact our team during project planning.
Bioinformatics Analysis
All FFPE spatial transcriptomics projects include a standard bioinformatics pipeline. Advanced and custom analyses are scoped during study design.
Standard Analysis (Included)
- Read alignment to probe reference and spatial barcode demultiplexing
- Feature-barcode expression matrix generation
- Spatial feature QC filtering and normalization
- Dimensionality reduction (PCA, UMAP)
- Unsupervised spatial clustering (graph-based, BayesSpace, or equivalent)
- Differential expression analysis between user-defined tissue regions or clusters
- Spatial feature plots — gene expression overlaid on H&E images
- Cell-type deconvolution using reference single-cell or single-nucleus data
- QC report with sequencing metrics, spatial feature statistics, and sample-level summaries
Optional Advanced Analysis
- Spatial neighborhood analysis — identification of recurrent cellular communities
- Ligand-receptor interaction inference incorporating spatial proximity
- Integration with single-nucleus RNA-seq (snRNA-seq) from matched FFPE blocks for enhanced cell-type resolution
- Multi-sample spatial data integration with batch correction
- Trajectory inference and spatially resolved gene program scoring
- Integration with spatial proteomics data from adjacent sections
- Comparison with fresh frozen spatial transcriptomics data from matched specimens
- Custom visualization — interactive HTML reports, spatial heatmaps, publication-quality figures
Analysis scope is agreed during study design to balance discovery depth against turnaround time. All data are delivered in formats compatible with common analysis environments (R, Python, Scanpy, Seurat, Giotto).
Deliverables
| Deliverable | Description |
|---|---|
| Raw sequencing data | Demultiplexed FASTQ files for all spatial barcoded libraries |
| Processed expression matrix | Feature × spatial barcode count matrix in standard formats (h5ad, h5, or mtx) |
| Spatial feature metadata | Tissue coordinates, in-tissue/background classification, and QC metrics per spatial feature |
| Spatial expression maps | Gene-by-gene spatial expression plots overlaid on H&E images (publication-ready PNG/SVG) |
| Clustering and annotation report | Spatial cluster assignments with marker gene tables and cluster annotation summaries |
| Cell-type deconvolution results | Estimated cell-type proportions per spatial feature, with reference dataset documentation |
| Differential expression tables | Differentially expressed genes between tissue regions or spatial clusters, with fold changes and adjusted p-values |
| Bioinformatics report | Methods documentation, QC metrics, analysis parameter logs, publication-ready figures, and interpretation notes |
| Data archive | All intermediate files, analysis scripts, and processing logs for reproducibility |
Applications
FFPE spatial transcriptomics unlocks archived tissue collections for spatially resolved gene expression analysis. It is most valuable when fresh tissue collection is impractical, when retrospective clinical cohorts with long-term follow-up are the primary research material, or when linking molecular data to histological features in standard pathology specimens.
Tumor microenvironment and cancer biology
Map gene expression across tumor core, invasive margin, and adjacent stroma in archived FFPE blocks. Identify spatially restricted gene programs associated with immune infiltration, tumor-stroma crosstalk, and treatment response. FFPE spatial transcriptomics has been applied to breast cancer, NSCLC, DLBCL, colorectal cancer, and cervical precancer specimens — revealing spatially organized immune cell populations and gene signatures that bulk RNA-seq cannot resolve.
Retrospective clinical cohort studies
Leverage FFPE blocks from completed clinical trials or longitudinal observational studies — some archived for 10+ years — with linked treatment and outcome data. Probe-based spatial transcriptomics adds a molecular spatial layer to existing histological annotations, enabling re-analysis of legacy cohorts with contemporary genomic tools.
Biomarker discovery and validation
Spatially resolved gene expression from FFPE specimens supports biomarker identification with direct histological context. Candidate biomarkers can be localized to specific tissue compartments — distinguishing tumor-cell-autonomous signals from stromal or immune-derived expression — before committing to orthogonal validation assays.
Immuno-oncology and immunotherapy research
Characterize immune cell populations, activation states, and spatial organization within FFPE tumor specimens. Probe-based spatial transcriptomics captures key immune genes (CD3, CD4, CD8, FOXP3, PD-L1, CTLA4, GZMB) alongside the full transcriptome, supporting integrated analysis of immune context and tumor biology from a single experiment. Pair with tumor microenvironment solutions for multi-assay study designs.
Neuroscience on archived brain specimens
Profile spatial gene expression in archived brain tissue from biobanks and brain banks. Map transcriptomic signatures across cortical layers, subcortical structures, and pathological lesions in FFPE specimens — linking molecular data to neuroanatomical features in specimens not originally collected for genomic analysis.
Drug development and preclinical pharmacology
Apply spatial transcriptomics to FFPE xenograft, PDX, or preclinical model tissues to assess how candidate compounds reshape gene expression across tissue architecture. Probe-based chemistry works consistently across human xenografts in mouse host tissue without species-specific protocol adjustments.
Case Study: Benchmarking Probe-Based Spatial Transcriptomics on Archived FFPE Tumors
Source: Dong Y, Saglietti C, Bayard Q, et al. Transcriptome analysis of archived tumors by Visium, GeoMx DSP, and Chromium reveals patient heterogeneity. Nature Communications 16, 4400 (2025).
Background: Archived FFPE tumor blocks represent a vast resource for translational research, but their fragmented RNA has historically limited spatial transcriptomic analysis. No systematic comparison of probe-based spatial transcriptomic methods had been performed on FFPE archival specimens — leaving uncertainty about which platform is best suited for cohort-scale studies and whether data quality from aged blocks matches that from freshly embedded specimens.
Methods: The study compared three probe-based whole-transcriptome methods on 16 FFPE tumor samples from 14 patients across three cancer types (breast carcinoma, non-small cell lung cancer, diffuse large B-cell lymphoma). Blocks had been archived for a median of 57 months (range up to 103 months). Visium CytAssist (probe-based spatial capture, ~18,000 genes), GeoMx Digital Spatial Profiler (WTA, ROI-based probe profiling), and Chromium Flex (probe-based single-nucleus RNA-seq) were all applied. Head-to-head comparisons assessed gene detection sensitivity, spatial resolution, tumor heterogeneity capture, and cell-type deconvolution accuracy.
Results: All three methods successfully generated transcriptome-wide data from archived FFPE material. Visium CytAssist captured unbiased spatial gene expression across whole tissue sections and, when integrated with Cell2location deconvolution, resolved spatially organized cell types that matched histological features. GeoMx DSP provided flexible ROI-level profiling but exhibited non-specific signal from cell mixtures within selected areas of interest — requiring careful marker-based interpretation. Chromium Flex achieved the highest cell-type resolution at single-nucleus level but lost all spatial context. The authors concluded that Visium and Chromium are most suitable for discovery of tumor heterogeneity, while GeoMx is best reserved for targeted questions on pre-specified regions.
Conclusion: Probe-based spatial transcriptomics generates high-quality whole-transcriptome data from FFPE blocks archived for nearly a decade. The choice of platform should be driven by the biological question — spatial discovery (Visium), single-cell resolution (Chromium Flex), or targeted ROI profiling (GeoMx) — and the study's tolerance for spatial context versus cell-type granularity.
FFPE Spatial Transcriptomics vs. Single-Nucleus RNA-seq
For FFPE specimens, two probe-based approaches can generate transcriptome-wide data. The key trade-off is spatial context versus cell-type resolution — understanding when each method is most informative helps plan a study design that captures both dimensions.
FFPE Spatial Transcriptomics vs. FFPE Single-Nucleus RNA-seq
| Factor | Spatial Transcriptomics (Visium FFPE) | Single-Nucleus RNA-seq (snRNA-seq) |
|---|---|---|
| Spatial context | Preserved — gene expression linked to tissue coordinates | Lost — nuclei are dissociated |
| Cell-type resolution | Deconvolved from multicellular features (or near-single-cell with HD) | Single-nucleus resolution |
| Gene detection | Probe-targeted; uniform coverage across gene body | Full-length transcript coverage with 5'/3' bias |
| Combined value | Best paired — snRNA-seq provides the single-nucleus reference for spatial deconvolution; spatial data adds tissue context to snRNA-seq clusters | |
Recommended strategy for archived specimens: Run spatial transcriptomics first for discovery. If single-nucleus resolution is required for cell-type annotation, add snRNA-seq from the same block. The two datasets can be integrated computationally to achieve spatially resolved single-nucleus-level gene expression maps.
Frequently Asked Questions (FAQ)
References
- Dong Y, Saglietti C, Bayard Q, et al. Transcriptome analysis of archived tumors by Visium, GeoMx DSP, and Chromium reveals patient heterogeneity. Nature Communications. 2025;16:4400. DOI: 10.1038/s41467-025-59005-9.
- Oliveira MF, Romero JP, Chung M, et al. High-definition spatial transcriptomic profiling of immune cell populations in colorectal cancer. Nature Genetics. 2025;57:1512–1523. DOI: 10.1038/s41588-025-02193-3.
- Madissoon E, Oliver AJ, Kleshchevnikov V, et al. A spatially resolved atlas of the human lung characterizes a gland-associated immune niche. Nature Genetics. 2023;55(1):66–77. DOI: 10.1038/s41588-022-01243-4.
- Muiños-Lopez E, Lopez-Perez AR, Sudupe L, et al. Characterization of the bone marrow architecture of multiple myeloma using spatial transcriptomics. Communications Biology. 2025;8:1620. DOI: 10.1038/s42003-025-08975-z.
- Castellanos M, Regan C, Khutti S, et al. Spatial transcriptomics for profiling the microenvironment of cervical precancer progression from routine H&E slides. Journal of Clinical Oncology. 2025;43(16_suppl):e17536. DOI: 10.1200/JCO.2025.43.16_suppl.e17536.
