Spatial transcriptomics is one of the most consequential additions to the drug discovery toolkit in the past decade, and its adoption is accelerating across oncology, neurodegeneration, and immunology pipelines. The technology resolves a problem that bulk RNA sequencing and single-cell RNA sequencing have never fully addressed: the loss of anatomical context. Understanding where a gene is expressed within a tissue, not just whether it is expressed, changes which targets you prioritise and which you abandon.
Why Tissue Context Changes Everything in Drug Discovery
Bulk RNA sequencing averages gene expression across thousands of heterogeneous cells, masking the signals that define disease biology at the cellular and microenvironmental level. Single-cell RNA sequencing, which profiles individual cells after tissue dissociation, recovers cellular identity but discards the spatial relationships between cells. Those relationships are not incidental. The functional behaviour of a cancer cell at a tumour invasion front differs from the same cell type in the tumour core, and that difference drives treatment response.
Drug targets identified without spatial context carry a specific failure risk: they may be expressed in the right cell type but in the wrong tissue compartment, or their activity may depend on ligand-receptor interactions with neighbouring cells that only exist in a specific anatomical niche. Spatial transcriptomics closes this gap by linking molecular state to anatomical position within intact tissue architecture.
What Is Spatial Transcriptomics?
Spatial transcriptomics is a molecular profiling technology that measures gene expression across tissue sections while preserving the spatial coordinates of each measurement. Unlike bulk or single-cell RNA sequencing, it maps where genes are expressed within intact tissue architecture, enabling researchers to link cellular identity with anatomical context. This positional information is what makes it directly applicable to drug discovery decisions that single-cell methods cannot fully inform.
How Spatial Transcriptomics Works: The Core Mechanism
Two dominant technical approaches define the current platform field. Sequencing-based methods, including 10x Genomics Visium and Slide-seq, capture transcriptomes at defined spatial coordinates by placing tissue sections onto arrayed capture probes. The output is a spatially resolved gene expression matrix that can be overlaid onto histological images. Imaging-based methods, including Vizgen MERSCOPE (which uses MERFISH technology) and Resolve Biosciences, detect individual RNA molecules in situ using multiplexed fluorescent probes, achieving subcellular resolution across hundreds of pre-selected genes.
The resolution and throughput trade-off between these approaches directly determines which drug discovery questions each platform can answer. Sequencing-based platforms offer whole-transcriptome coverage at lower per-gene cost but typically at multicellular spot resolution rather than single-cell precision. Imaging-based platforms achieve subcellular resolution and are better suited to hypothesis-driven spatial profiling of known pathways, but require gene panel pre-selection and carry higher per-sample costs at scale.
Spatial Transcriptomics vs. Single-Cell RNA Sequencing: Key Differences
| Dimension | Bulk RNA-seq | Single-Cell RNA-seq | Spatial Transcriptomics |
|---|---|---|---|
| Spatial resolution | None | None (dissociated) | Multicellular to subcellular |
| Gene coverage | Whole transcriptome | Whole transcriptome | Whole or targeted panel |
| Microenvironment mapping | No | Partial (inferred) | Direct |
| Best drug discovery use | Expression profiling | Cell-type identification | Target ID, MOA, TME analysis |
Target Identification: Moving Beyond the Bulk Transcriptome
Spatial transcriptomics reveals cell-type-specific expression patterns within disease-relevant tissue regions that bulk sequencing averages out entirely. In oncology, this means you can profile gene expression at the tumour invasion front separately from the hypoxic core, or isolate the transcriptomic signature of cancer-associated fibroblasts within a fibrotic niche. Each of these spatial zones represents a distinct biological state with distinct target opportunities.
Targets identified in spatially defined disease zones carry a stronger mechanistic rationale from the outset. You’re not asking whether a gene is expressed in tumour tissue generally; you’re asking whether it’s expressed in the specific cellular neighbourhood where disease progression is occurring. That distinction matters for patient stratification hypotheses and for the IND-enabling studies that follow. Published research, including work by Cao and colleagues in 2024, positions spatial transcriptomics as a direct input to novel target discovery workflows rather than a post-hoc validation tool.
How Spatial Transcriptomics Is Used for Drug Target Identification
- Tissue collection and preparation: Fresh-frozen or FFPE tissue sections are prepared to preserve RNA integrity and spatial architecture.
- Spatial library generation: Sections are placed on capture arrays or processed for in situ hybridisation, depending on the platform.
- Sequencing and image alignment: Gene expression data is computationally aligned to histological images using tissue landmarks.
- Spatially variable gene identification: Statistical methods identify genes whose expression varies significantly across tissue coordinates.
- Cell-type deconvolution: Computational tools disaggregate multicellular spots into constituent cell types using reference single-cell atlases.
- Target prioritisation: Candidates are ranked by spatial specificity, expression magnitude in disease-relevant zones, and druggability.
- Validation strategy: Prioritised targets are validated using orthogonal methods, including multiplexed immunofluorescence or in situ sequencing.
Mapping Drug Mechanism of Action at Tissue Resolution
Spatial transcriptomics can profile how a drug perturbs gene expression across different tissue compartments simultaneously. This is where the technology earns its place in translational workflows rather than just discovery. A compound that suppresses a target in vitro may show a completely different activity profile when applied to a complex tissue environment where stromal cells, immune infiltrates, and vasculature all modulate drug access and signalling.
Intra-tumoral heterogeneity, the coexistence of drug-sensitive and drug-resistant cell populations within the same tumour, is directly visible in spatial data. Resistance mechanisms often emerge from specific spatial niches. Identifying those niches in preclinical tissue models before IND-enabling studies gives your programme an opportunity to design combination strategies or patient selection criteria that address resistance before it becomes a Phase II problem.
The practical implication for your pipeline: mechanism-of-action mapping with spatial resolution is most valuable for programmes where in vitro efficacy hasn’t translated cleanly into in vivo models. If a compound works in cell lines but fails in complex tissue, spatial transcriptomics is one of the few tools that can tell you why with the mechanistic specificity needed to act on that information.
Tumour Microenvironment and Resistance Profiling
The tumour microenvironment, comprising immune cells, stromal cells, vasculature, and cancer cells, is spatially organised. That organisation determines treatment response. Immune cell localisation relative to tumour cells is a direct predictor of immunotherapy response, and spatial transcriptomics makes that localisation quantifiable at the transcriptomic level rather than relying on protein markers alone.
Spatial biology enables precise mapping of cell-cell communication through ligand-receptor interaction analysis. You can identify which cancer cells are receiving immunosuppressive signals from adjacent myeloid populations, and where in the tissue that communication is most active. This directly informs immuno-oncology target selection. The question isn’t just whether an immune checkpoint is expressed; it’s whether it’s expressed at the interface between tumour and immune cells where it would actually mediate suppression.
Platform Selection: Matching Technology to the Drug Discovery Question
Platform choice has downstream consequences that extend well beyond the experiment itself. The data volume, computational infrastructure requirements, and the type of IP claims you can build on spatial data all depend on which platform generated it. For early-stage biotech founders, this is a build-vs-buy decision with long-term implications for data ownership and regulatory filing strategy.
10x Genomics Visium remains the most widely adopted sequencing-based platform, offering whole-transcriptome coverage at a resolution of approximately 55 microns per spot. Nanostring CosMx provides single-cell resolution imaging with panels covering hundreds to thousands of genes, making it well-suited for detailed tumour microenvironment profiling. Vizgen MERSCOPE achieves subcellular resolution using MERFISH chemistry, which is particularly valuable for neuroscience and CNS drug discovery applications where cell-type boundaries are architecturally significant.
When evaluating CRO partnerships for spatial profiling work, you need clarity on three points: who owns the raw data files, what computational pipeline will be applied, and whether the resulting biomarker claims are defensible under your IP strategy. Those terms matter as much as the biology.
AI and Computational Integration in Spatial Biology Pipelines
Raw spatial transcriptomics datasets are structurally complex. Extracting drug discovery signal requires computational pipelines that integrate spatial statistics, cell-type deconvolution, and pathway analysis into a coherent analytical output. Deconvolution, the process of disaggregating mixed-cell expression signals into constituent cell-type contributions, is computationally demanding and depends heavily on the quality of the reference single-cell atlas used.
Published frameworks, including work cited in the 2024 computational biology literature, propose integrating spatial transcriptomics with computer vision techniques to automate therapeutic target identification from tissue images. AI-assisted spatial analysis is moving from academic proof-of-concept toward platform-level integration at CROs and larger biotech companies. For your programme, this means data infrastructure decisions made now will affect your ability to integrate spatial data with proteomics, metabolomics, and clinical imaging data later.
Strategic Implications for Biotech Founders and Drug Discovery Teams
Spatial transcriptomics data generated during target identification can anchor IP claims around spatially defined biomarkers and patient stratification methods. This is an underexploited filing opportunity. A claim that a target is expressed in a specific spatial zone of disease tissue, with a defined cellular neighbourhood, is more defensible and more clinically actionable than a claim based on bulk expression data alone.
Early integration of spatial profiling into preclinical workflows reduces late-stage attrition risk by surfacing microenvironmental liabilities before IND-enabling studies begin. The technology is mature enough in oncology and immunology applications to justify that investment. In CNS drug discovery and spatial toxicology, the evidence base is earlier-stage, and you should weight platform decisions accordingly.
The spatial biology field has grown substantially in published output since 2020, and the competitive pressure to integrate these methods is real. The programmes that will benefit most are those that treat spatial transcriptomics as a pipeline decision tool rather than a discovery curiosity, connecting tissue-level data to target prioritisation, biomarker strategy, and trial design from the earliest stages of development.
Frequently Asked Questions
How does spatial transcriptomics improve target identification compared to single-cell RNA sequencing alone?
Spatial transcriptomics adds anatomical context that single-cell RNA sequencing loses during tissue dissociation. By preserving the physical location of gene expression within intact tissue, it identifies targets that are selectively expressed in disease-relevant spatial zones, such as tumour invasion fronts or fibrotic niches, providing a stronger mechanistic rationale and clearer patient stratification hypothesis from the outset.
What are the key limitations of spatial transcriptomics in drug discovery?
Current limitations include throughput constraints at single-cell resolution, tissue processing artifacts that can degrade RNA quality in FFPE samples, and computational scalability challenges when integrating spatial data with other omic layers. Sequencing-based platforms also face resolution trade-offs that limit single-cell precision without computational deconvolution.
Which spatial transcriptomics platforms are used in drug discovery?
The most widely used platforms in drug discovery include 10x Genomics Visium for whole-transcriptome sequencing-based spatial profiling, Nanostring CosMx for single-cell resolution imaging, and Vizgen MERSCOPE for subcellular MERFISH-based profiling. Platform selection depends on the specific research question, required resolution, gene panel requirements, and available computational infrastructure.
