Spatial Omics Solutions for Drug Discovery Research
Promising targets and treatment effects can disappear in tissue-level averages when activity is confined to specific cell states, disease compartments, or microenvironmental niches. Spatial omics helps reveal where a target is expressed, which regions respond to treatment, where resistant or escape-like programs persist, and whether local drug distribution aligns with molecular response. CD Genomics integrates single-cell discovery, spatial transcriptomics, spatial proteomics, and optional mass spectrometry imaging into a study plan matched to your model, tissue format, treatment comparison, and development question. The resulting evidence can support target prioritization, mechanism-of-action research, biomarker exploration, combination-strategy hypotheses, and selection of the most informative next experiment.
Tissue-Relevant Target Evidence
Regional Mechanism of Action
Heterogeneous Response and Resistance
Why Drug-Discovery Signals Become Ambiguous in Heterogeneous Tissue
A target can be highly expressed in the wrong compartment, and a treatment can change a whole-tissue average while leaving a protected region unaffected. Spatial omics is most useful when the unresolved question depends on where a signal occurs, which cells carry it, and what local tissue context may enable or restrict the effect.
1
Abundance does not establish target relevance
Bulk profiling can rank a transcript or protein highly because it is widespread in a large background compartment. Without localization, the team cannot tell whether the candidate is present in the disease-driving niche, an unrelated cell population, or both.
2
Response does not automatically explain mechanism
A pathway score may confirm that treatment changed the tissue, but it does not identify the responding compartment, distinguish a cell-state shift from a composition shift, or prove direct target engagement. The mechanism remains underdetermined until the signal is anchored to tissue context and an appropriate validation step.
3
Local resistance disappears in the average
Strong-response, weak-response, escape-like, and remodeling regions can coexist within one treated section. Pooling them into a single value can dilute the region that determines the next experiment, while post hoc hotspot selection can mistake necrosis or technical failure for non-response.
Decision rule: If the uncertainty is global—such as whether a treatment has any overall effect—bulk profiling may remain the most efficient first step. If the uncertainty is local—such as which compartment responds, where a candidate is expressed, or why adjacent regions diverge—spatial profiling can change the interpretation.
Where Spatial Omics Changes a Drug-Discovery Decision
We structure each project around the drug-development decision that must be settled: whether a target belongs in the relevant tissue niche, how a treatment acts across compartments, or why response differs between adjacent regions. The evidence plan then uses only the measurements needed to support that decision.
Target
Target Discovery and Prioritization
Determine whether a candidate is expressed in a decision-relevant region, carried by a disease-associated cell state, and supported by a plausible local biological context.
- Location: Is the candidate present in the disease-driving niche or only in background tissue?
- State: Which cell state carries the signal, and is that state enriched in disease?
- Follow-up: Is there a realistic perturbational or protein-level validation path?
MOA
Tissue Mechanism of Action
Compare treated and reference tissue to locate pathway shifts, state transitions, neighborhood remodeling, and—when measured—compound distribution.
- Compartment: Where does the earliest or strongest response occur?
- Mechanistic hinge: Which pathway, cell state, or interaction changes with treatment?
- Boundary: Which result is measured, which is inferred, and what requires direct validation?
Response
Regional Response and Resistance
Replace a single responder/non-responder label with reproducible regional classes that remain connected to morphology and whole-section context.
- Definition: Lock the biological response criterion before selecting regions.
- Contrast: Compare strong-response, weak-response, transition, and non-informative areas.
- Next step: Rank spatial hypotheses that can be tested in a follow-up experiment.
How to Match Spatial and Single-Cell Technologies to the Drug Question
Technology selection should follow the uncertainty the study must reduce. Broad discovery, cell-level localization, protein confirmation, perturbational testing, and compound imaging answer different questions; combining them is useful only when each layer changes the interpretation or the next decision.
| Technology | Analytical Role | Useful Output | When It Adds Value | Boundary |
|---|---|---|---|---|
| scRNA-seq / snRNA-seq 10x Genomics Chromium |
Discover treatment-associated cell states, rare populations, and candidate markers | Cell-state atlas, UMAP, marker matrix, differential-state results | Before spatial profiling, or when an existing single-cell reference can guide tissue mapping | Dissociation or nuclei isolation removes tissue position; findings require spatial localization when location matters |
| Perturbational single-cell sequencing CROP-seq |
Connect selected genetic perturbations with transcriptional phenotypes | Guide-linked cell states, perturbation effects, candidate-response programs | When a shortlist needs functional prioritization in a suitable cellular model | Model-dependent and not spatial by itself; it does not establish druggability or in-tissue efficacy |
| Whole-transcriptome spatial profiling Visium, Visium HD, Stereo-seq |
Map broad gene programs, tissue domains, and response architecture | Spatial expression map, tissue-domain analysis, pathway and cell-state localization | Hypothesis-light discovery or whole-section comparison across treatment groups | Resolution and cell assignment vary by platform; expression shifts do not prove direct target engagement |
| Targeted in situ RNA profiling Xenium, CosMx SMI |
Localize selected transcripts at cell or subcellular scale | Cell-segmented target maps, neighborhood composition, boundary localization | When targets or signatures are already shortlisted and precise localization is decisive | Panel-limited; it is not a substitute for hypothesis-light whole-transcriptome discovery |
| Spatial proteomics GeoMx DSP, mIHC |
Evaluate selected protein phenotypes, pathway markers, and morphology-defined regions | Protein expression matrix, spatial phenotype map, RNA–protein comparison | When RNA-defined findings need an orthogonal protein evidence layer | Panel-limited and associative; protein abundance alone does not prove functional causality |
| Spatial metabolomics / molecular imaging MALDI-MSI, DESI-MSI |
Map compounds, lipids, metabolites, and regional chemical context | Ion-distribution maps, region comparisons, registered molecular overlays | When compound localization or metabolic response may explain regional efficacy | Requires identification, normalization, registration, and orthogonal biological interpretation |
Integration logic: Single-cell data can identify which states changed; spatial transcriptomics can show where those states and pathways occur; spatial proteomics can test selected protein phenotypes; and MSI can show where a compound or metabolite is distributed. A combined result is strongest when the assays share an explicit sample-registration, comparison, and validation plan.
If you already have single-cell data, the single-cell + spatial transcriptome integration route can use it as a reference instead of repeating the discovery step.
Study Designs for Target, MOA, and Regional-Response Questions
These are configurable study routes, not fixed commercial tiers. Final platforms, sample numbers, section strategy, and analysis depth are set after review of tissue preservation, biological replication, treatment groups, timepoints, morphology, and the decision the study must support.
Target Prioritization Route
For teams deciding which tissue-relevant candidates should move into functional follow-up
Comparison Logic
Disease vs reference tissue, target-high vs target-low regions, or disease-driving vs background compartments
Evidence Layers
Single-cell or nuclei reference + whole-transcriptome spatial discovery; optional targeted RNA or protein confirmation
Key Deliverables
Candidate-location map, cell-state attribution, neighborhood context, prioritization matrix, and follow-up queue
Evidence Boundary
Proof of druggability, direct binding, causal function, or treatment efficacy
Mechanism-Focused
Tissue MOA Route
For teams that observe activity but need to localize the responding compartment and mechanism
Comparison Logic
Treated vs control, dose series, early vs later exposure, or target-rich vs target-poor regions
Evidence Layers
Spatial transcriptomics + selected spatial protein markers; optional MSI when compound distribution is part of the question
Key Deliverables
Spatial pathway map, state-shift analysis, neighborhood remodeling, distribution–response overlay, and ranked mechanism hypotheses
Evidence Boundary
Direct target-binding or causal proof unless separately scoped with an appropriate validation assay
Regional Response Route
For preclinical tissues or research cohorts with mixed response, resistance, or escape-like regions
Comparison Logic
Pre-defined strong-response, weak-response, transition, escape-like, and non-informative regions within whole-section context
Evidence Layers
Whole-section spatial RNA + morphology review; optional protein validation and cell-neighborhood analysis
Key Deliverables
Region classification map, replicate-aware contrasts, candidate resistance programs, validation priorities, and QC exclusions
Evidence Boundary
Clinical response prediction, individual treatment selection, or universal responder biomarkers
What to prepare for design review: tissue type and preservation; biological replicate structure; treatment, dose, and timepoint groups; the expected responding and compensatory compartments; a one-sentence response definition; and the decision that would change if the spatial hypothesis is supported or rejected.
Replication boundary: Spots, cells, or regions from one section are not independent biological replicates. Statistical inference should follow the true biological unit—such as animal, donor, organoid, or independently prepared model—and account for paired, longitudinal, or nested designs where applicable.
Analysis Outputs Organized Around the Next Drug-Development Decision
The goal is not to deliver more plots. Outputs are grouped so a discovery team can decide which target, mechanism, response pattern, or validation experiment should move forward. Exact figures and file formats depend on the selected platforms and agreed analysis scope.
Target-Prioritization Matrix
Ranks candidates by location, carrier cell state, disease association, neighborhood plausibility, evidence quality, and available validation path. This helps separate a broadly expressed marker from a candidate positioned in the relevant tissue niche.
Treatment-Response Tissue Map
Overlays response-associated states and pathways on tissue morphology, with strong-response, weak-response, transition, escape-like, and non-informative regions clearly distinguished.
State and Neighborhood Shift Analysis
Separates changes in cell abundance from changes within a cell state, then evaluates whether local communities, boundaries, or candidate interactions shift with treatment.
Mechanism and Validation Queue
Labels each finding as measured evidence, computational inference, or unresolved hypothesis, then recommends an appropriate next step such as targeted RNA mapping, protein staining, perturbation, or a refined model design.
Target, State, and Regional-Response Dashboard
A coordinated view of candidate location, cell-state attribution, spatial pathway activity, treatment contrasts, and evidence status for multidisciplinary review.
Compound Distribution–Response Overlay
A conceptual co-registration of molecular imaging, tissue morphology, and spatial response programs to ask whether local exposure and local biology agree.
Research Interpretation Note: A spatial association can prioritize a mechanism, but it does not establish causality. Ligand–receptor analysis predicts compatible signaling relationships; it does not directly measure physical interaction. When communication is central to the hypothesis, consider a dedicated intercellular communication analysis followed by targeted validation.
Why Choose CD Genomics for Spatial Drug-Discovery Research
Single-Cell + Spatial Integration Workflow
Single-cell or nuclei reference states can be mapped into Visium, Visium HD, Stereo-seq, Xenium, or CosMx data with the mapping assumptions and confidence retained in the analysis. This lets an existing cell atlas guide tissue localization instead of becoming a disconnected dataset.
Multi-Modal Spatial Assay Selection
Transcript, protein, and metabolite-level options are selected according to coverage, resolution, tissue area, preservation, cohort scale, and the decision the output must support. The study can stay focused rather than adding modalities that do not change the next step.
FFPE and Fresh-Frozen Study Routes
FFPE-compatible spatial transcriptomic, in situ RNA, and spatial protein options can support archived-tissue studies, while fresh or frozen material can enable broader single-cell and whole-transcriptome routes. Final feasibility depends on fixation history, analyte quality, morphology, section area, and platform-specific QC.
Decision-Led Custom Bioinformatics
Analyses can be organized around target ranking, treated-versus-reference contrasts, region classification, cell-state mapping, neighborhood changes, and evidence boundaries. The team receives interpretable outputs tied to the research decision rather than a generic list of available algorithms.
For pathway-centered studies, review the Spatial Omics Solutions for Pathway Analysis. For tissue safety and organ injury questions, use the Spatial Omics Solutions for Toxicology. For tumor–stroma and immune-niche mechanisms, see the Tumor Microenvironment solution.
Published Research Examples
These independent published studies illustrate two spatial drug-discovery questions: how local compound accumulation relates to molecular response, and how tissue regions can differ in inferred drug sensitivity. The illustrations are original conceptual summaries, not reproductions or adaptations of published figures.
Example 1: Linking PARP-Inhibitor Distribution to Regional Molecular Response
Moncayo CR, Restuadi R, Zhang G, et al. Nature Communications. 2026;17:4086. doi:10.1038/s41467-026-70558-1
Research Question
Does intratumoral heterogeneity affect the local accumulation of PARP inhibitors, and can spatial molecular profiling explain how drug-rich and drug-poor regions differ?
Study Design
Patient-derived explants from high-grade serous ovarian carcinoma were treated ex vivo with PARP inhibitors. The study combined mass spectrometry imaging of drug distribution, GeoMx whole-transcriptome spatial profiling on adjacent sections, immunohistochemistry, and cell-level follow-up experiments.
Key Findings
- PARP-inhibitor accumulation varied between patients, between tumors, and within individual tumor samples.
- Regions with higher niraparib or rucaparib signal showed stronger apoptotic-response programs and lysosome-associated transcriptional signatures.
- Cell-level experiments supported a role for lysosomal sequestration in heterogeneous rucaparib and niraparib accumulation.
Relevance to This Solution
The study illustrates why compound imaging and spatial transcriptomics answer different parts of a pharmacology question. One layer mapped where drug accumulated; another connected those regions to molecular response, helping separate exposure heterogeneity from response biology.
Boundary
The spatial analysis used adjacent sections from a small patient-derived explant cohort, and the findings concern selected PARP inhibitors in this ovarian-cancer context. Co-registration, tissue differences between sections, and independent validation remain important design considerations.
Example 2: Spatial Profiling Reveals Regional Drug-Response Heterogeneity in Breast Cancer
Jiménez-Santos MJ, García-Martín S, Rubio-Fernández M, et al. NAR Cancer. 2024;6(4):zcae046. doi:10.1093/narcan/zcae046
Research Question
Can spatial transcriptomic profiles distinguish tumor and microenvironment regions with different predicted sensitivities to anticancer compounds?
Study Design
The researchers analyzed breast-cancer spatial transcriptomic datasets, separated tumor and tumor-microenvironment regions, and applied computational drug-sensitivity inference to compare regional therapeutic-response patterns.
Key Findings
- Tumor and microenvironment regions showed distinct transcriptional programs relevant to therapeutic response.
- Computational inference identified regional gradients and heterogeneity in predicted drug sensitivity.
- The analysis linked candidate treatment responses to the spatial organization of malignant and surrounding tissue compartments.
Relevance to This Solution
The study shows why a tissue-average response estimate can miss region-specific behavior. Spatially resolved analysis can help prioritize which compartments, response programs, or candidate drugs warrant experimental follow-up.
Boundary
Drug sensitivity was computationally inferred from spatial expression data rather than prospectively measured as treatment outcome. The results support hypothesis generation and experimental prioritization, not clinical treatment selection.
Frequently Asked Questions
Spatial omics adds the most value when the unresolved question depends on location: whether a target sits in the disease-driving compartment, which region responds first, why adjacent regions diverge, whether a stromal or immune neighborhood supports resistance, or whether compound distribution aligns with molecular response. If the decision is still a global ranking of treatments or confirmation of an overall effect, a bulk assay may be the more efficient first step.
Not by itself. Spatial transcriptomics can localize pathway changes, cell-state transitions, and regional response, but expression data are usually associative. Direct target engagement, binding, causal function, or functional rescue requires an assay appropriate to that claim. A strong study states which result is measured, which is computationally inferred, and which follow-up experiment would test the mechanism.
Define response before selecting regions. The definition may combine morphology, a pharmacodynamic marker, pathway activity, cell-state change, or another pre-specified endpoint. Region classes should be reproducible across matched sections or biological replicates, while necrosis, folds, tears, low cellularity, and assay-QC failures should be labeled non-informative rather than interpreted as biological non-response.
Yes, if the reference is biologically relevant and its metadata and quality are sufficient. Existing cell-state signatures can guide deconvolution or label transfer into new spatial data, but mapping confidence depends on platform resolution, tissue match, state representation, batch effects, and the integration method. Results should distinguish measured spatial expression from computationally mapped cell identities.
FFPE-compatible options include whole-transcriptome spatial assays, targeted in situ RNA profiling, GeoMx DSP, and multiplex protein imaging. Suitability depends on fixation and storage history, analyte quality, tissue area, morphology, target abundance, cohort scale, and platform-specific QC. A feasibility review should precede final sample allocation or panel design.
There is no universal number. The design depends on the biological unit, treatment structure, expected heterogeneity, paired or longitudinal sampling, model-to-model variation, tissue availability, and planned statistical contrast. Thousands of spots or cells from one section do not replace independent biological replicates. A pilot can establish assay feasibility, but confirmatory sample size requires study-specific statistical review.
Provide tissue type and preservation, species or model, section or block availability, treatment groups, doses and timepoints, biological replicate structure, pathology or morphology information, existing single-cell or bulk data, the expected responding and compensatory compartments, and the decision the study should change. These inputs are more useful than naming a preferred platform before the study question is defined.
Related Services, Solutions, and Study-Design Resources
Continue from your drug-development question to a relevant spatial service, adjacent solution, or study-design resource.
References
Moncayo CR, Restuadi R, Zhang G, et al. Multimodal imaging reveals a lysosomal drug reservoir that drives heterogeneous distribution of PARP inhibitors. Nature Communications. 2026;17:4086. https://doi.org/10.1038/s41467-026-70558-1
Jiménez-Santos MJ, García-Martín S, Rubio-Fernández M, et al. Spatial transcriptomics in breast cancer reveals tumour microenvironment-driven drug responses and clonal therapeutic heterogeneity. NAR Cancer. 2024;6(4):zcae046. https://doi.org/10.1093/narcan/zcae046
For research purposes only. Not intended for clinical diagnosis, treatment, or individual health assessment.
Online Inquiry