Case Study
Single-Cell and Spatial Omics in Drug Discovery

Inquiry
Single-Cell and Spatial Omics in Drug Discovery - CD ComputaBio
Opening perspective

Seeing Disease Beyond the Average

Single-cell and spatial omics provide complementary ways to examine disease biology at resolutions that bulk measurements cannot preserve. Single-cell profiling resolves which cell populations and molecular states are present, while spatial approaches retain information about where those cells or molecular signals occur within intact tissue.

For drug discovery, the value is not simply higher resolution. These data can separate a disease-driving cellular program from a signal diluted across an entire specimen, localize a candidate target to a pathological niche, reveal which populations respond to treatment, and generate biomarker hypotheses based on cell states or tissue organization. The analytical objective is to convert high-dimensional observations into evidence that can be prioritized, reproduced, and tested.

Neither modality establishes therapeutic causality on its own. Cell-type-specific expression, spatial proximity, pathway enrichment, and predicted communication are candidate evidence. Confidence grows when the signal replicates across donors, remains consistent across modalities or cohorts, and can be examined using orthogonal molecular or functional experiments.

Cell identityWhich populations are present?
Cell stateWhat programs are active?
Tissue locationWhere do signals occur?
Therapeutic responseWhich populations change?
Complementary resolution

From Cellular Heterogeneity to Tissue Context

Dissociated single-cell measurements and tissue-resolved assays answer different parts of the same biological question.

What single-cell omics resolves

Single-cell RNA sequencing and related single-cell modalities distinguish major populations, rare subsets, transitional states, lineage-associated programs, and cell-type-specific responses that may be obscured in a sample-level average. With appropriate biological replication, researchers can compare how a defined cell population changes across disease, treatment, time, genotype, or model system.

Multiomic designs can connect transcript abundance with chromatin accessibility, surface proteins, immune receptors, genetic perturbations, or other measurements. These combinations can strengthen a regulatory hypothesis, but integration quality depends on assay compatibility, feature coverage, missingness, and whether modalities are measured in the same cells or aligned computationally.

What spatial omics adds

Spatial transcriptomic, proteomic, and emerging multimodal approaches retain molecular measurements within tissue architecture. Depending on platform, a measurement may represent a multicellular spot, an inferred cell, a segmented cell, or a subcellular location. This distinction matters when interpreting spatial domains, cellular neighborhoods, and putative interactions.

Spatial information can show whether a disease-associated state is concentrated at an invasive boundary, fibrotic region, immune aggregate, vascular compartment, or other pathological niche. It can also test whether a candidate signal is consistently localized across sections and individuals rather than being scattered throughout tissue.

Together, the two views are complementary: single-cell data primarily establishes which cells and states are present; spatial data establishes where molecular programs and cellular neighborhoods are organized. Joint interpretation links identity with location, but does not remove the need for replication or experimental validation.
Signature component

A Technology Landscape for High-Resolution Biology

The field can be organized along two dimensions: preservation of tissue context and the number of molecular modalities being profiled or integrated.

Dissociated cellular profilesIntact tissue context
Single modality → Integrated multiomics
Single-cell mono-omics

High-resolution measurements of one principal molecular layer in dissociated cells.

scRNA-seqscATAC-seqImmune profiling
Spatial mono-omics

One molecular layer mapped within tissue sections at platform-dependent resolution.

Spatial transcriptomicsSpatial proteomics
Single-cell multiomics

Joint or aligned cell-level modalities for regulatory and perturbational analysis.

RNA + chromatinRNA + proteinPerturbation readouts
Spatial multiomics

Multiple molecular layers analyzed with coordinates, images, or tissue regions.

Spatial RNA + proteinSpatial epigenomics
Cellular and tissue-resolved disease model
Quality control
Data integration
Cell states
Spatial niches
Evidence prioritization
Single-cell and spatial omics span complementary experimental and analytical dimensions, from dissociated cellular profiles to multimodal measurements retained within tissue architecture.
Analytical framework

Turning High-Dimensional Data into Biological Evidence

Reliable interpretation begins before clustering or visualization. Each analytical layer should preserve the study design and make uncertainty visible.

1

Data and metadata integrity

Confirm sample provenance, biological replicates, disease and treatment labels, batch structure, tissue regions, reference genome, feature annotations, and clinical covariates.

2

Cell- and location-level QC

Evaluate low-quality cells or spots, doublets, ambient RNA, sparsity, imaging artifacts, segmentation, deconvolution, and platform-specific resolution.

3

Representation and integration

Apply fit-for-purpose normalization, batch-aware integration, cross-sample alignment, reference mapping, and single-cell–spatial integration without erasing biological differences.

4

Biological interpretation

Resolve cell identities and states, differential programs, pathways, spatial domains, neighborhoods, regulatory candidates, and treatment-associated changes.

5

Evidence prioritization

Rank findings by donor replication, effect size, uncertainty, modality consistency, disease specificity, external evidence, and experimental testability.

The analysis should respect the unit of replication. Thousands of cells from one sample do not replace independent biological samples. For condition-level inference, donor- or sample-aware models are generally needed to avoid pseudoreplication. Batch correction also cannot recover a biological contrast that is completely confounded with processing batch. These design constraints should be identified during the metadata audit rather than hidden at the reporting stage.

Signature component

Where Single-Cell and Spatial Omics Inform the Drug Development Continuum

High-resolution omics can contribute different evidence at each stage, from mapping disease ecosystems to developing testable biomarkers.

Stage 01

Disease mapping

Identify disease-associated cell populations, cell states, spatial regions, and abnormal cellular ecosystems.

Output: disease cell atlas
Stage 02

Target discovery

Evaluate cell-type-specific expression, dysregulated programs, spatial restriction, and links to disease phenotypes.

Output: prioritized target hypotheses
Stage 03

Candidate and MOA studies

Characterize perturbation-associated state changes, responsive populations, and pathways consistent with a proposed mechanism.

Output: MOA hypotheses
Stage 04

Preclinical response and safety

Map tissue-specific responses, unanticipated cellular effects, and cell states associated with efficacy or toxicity in model systems.

Output: response and safety signals
Stage 05

Biomarkers and stratification

Associate cellular composition, state signatures, or spatial features with treatment response, outcome, or molecular subtype.

Output: candidate biomarkers
Interpretation boundary: cell-specific expression does not by itself establish causality or druggability; predicted ligand–receptor relationships do not demonstrate physical signaling; treatment-associated changes require appropriate controls; and predictive biomarkers require patient-level modeling plus independent validation.
Fit-for-purpose design

Matching the Omics Strategy to the Research Question

A technology should be selected for the decision it must support, not simply because it offers greater nominal resolution.

Research questionUseful evidenceImportant qualification
Which population carries the disease signal?Single-cell profiling across biologically replicated groupsPopulation abundance and state changes require sample-aware inference.
Where is the disease-associated population located?Spatial profiling with histological contextResolution and segmentation determine whether a signal is cell-specific.
Which regulatory programs distinguish the population?Single-cell transcriptomic or multimodal analysisRegulatory links are hypotheses unless supported by perturbation or orthogonal data.
Is a candidate signal restricted to a pathological niche?Spatial mapping across sections and individualsSection selection and tissue heterogeneity affect generalizability.
How does treatment alter heterogeneous cell states?Longitudinal, paired, or perturbational single-cell analysisTime, dose, model, and baseline state must be represented in the design.
Does a signature associate with patient outcome?Single-cell discovery followed by patient-level cohort validationPerformance must be assessed outside the discovery data.

The practical choice also depends on tissue availability, preservation, cell viability, cohort size, required molecular modality, imaging quality, acceptable resolution, and the feasibility of downstream validation. In some programs, bulk profiling remains the most efficient approach for a large validation cohort, while single-cell or spatial data serve as a focused discovery layer.

Interpretation risks

Challenges That Shape the Strength of a Conclusion

The most informative analysis is one that makes design limitations explicit and prevents technical structure from being interpreted as disease biology.

Biological sampling

Tissue dissociation can preferentially lose fragile populations or induce stress programs. Ischemic time, processing delay, anatomical region, disease stage, medication, and donor composition can all alter the observed cellular landscape.

Technical measurement

Sparse counts, ambient RNA, doublets, low-complexity cells, spot mixing, image quality, segmentation, and deconvolution introduce different forms of uncertainty. Quality criteria should be assay- and tissue-aware.

Statistical design

Cells are nested within samples and are not independent replicates. Batch–condition confounding, unequal sample sizes, multiple testing, compositional effects, and selective reporting can inflate apparent evidence.

Translation

Cell states and spatial features may shift across cohorts, platforms, disease stages, and sample handling procedures. Candidate findings require reproducibility, assay simplification, and orthogonal or functional validation.

Open-access evidence

Single-Cell Multiomics Connects Molecular Layers with Drug Research Questions

A 2024 open-access review describes how single-cell multiomics can jointly interrogate genomic, epigenomic, transcriptomic, and protein-level features in complex disease and drug-response research. Its framework illustrates the potential to connect cell heterogeneity with drug–target binding and response mechanisms.

The figure is included as evidence of the technology landscape, not as proof that every multimodal assay is required or that computational integration establishes mechanism. Study-specific conclusions still depend on appropriate biological models, replicate structure, assay coverage, statistical analysis, and experimental confirmation.

Spatial approaches add a further dimension by placing molecular signals in tissue context. Reviews of spatial multiomics emphasize both the opportunity for target and biomarker research and the need for standardized experimental design, integration, and interpretation.

Single-cell multiomics strategies for studying complex disease and mapping drug-target interactions
Single-cell multiomics strategies for studying complex disease and mapping drug–target relationships.1
Practical takeaways

Questions to Resolve Before Starting a Study

A focused design begins with the biological decision and works backward to the samples, assays, analysis, and validation needed to support it.

What biological or development decision should the analysis support?
What is the independent biological replicate: donor, animal, organoid, or culture?
Is tissue dissociation acceptable, or is native location essential?
Which metadata, treatments, time points, and covariates are available?
Can biological condition be separated from processing batch?
What resolution and molecular modality are truly needed?
What evidence will qualify a finding for follow-up?
Which external dataset, orthogonal assay, or functional experiment can validate it?
Bioinformatics support

Connecting High-Resolution Data to Research Decisions

ComputaBio can help define an analysis strategy around the biological question, available data, replicate structure, and intended validation path.

Single-Cell RNA-Seq Data Analysis

Cell- and sample-level quality control, integration, annotation, differential state analysis, and biologically grounded interpretation.

Explore bioinformatics support →

Target Discovery from Omics Data

Integration of cell-specific, pathway, disease, and external evidence to generate transparent target-prioritization hypotheses.

Discuss target discovery →

Biomarker Discovery from Omics Data

Candidate cellular or molecular signature development with attention to patient-level inference and independent validation.

Discuss biomarker analysis →

Discuss Your Study Design

Share your biological question, sample structure, available data, and intended decision point. We can help define a fit-for-purpose single-cell or spatial omics analysis plan with explicit quality, interpretation, and validation criteria.

Request a Customized Analysis Plan
References

Selected Scientific Sources

  1. Ma, J.; Dong, C.; He, A.; Xiong, H. Single-cell multiomics: a new frontier in drug research and development. Frontiers in Drug Discovery 2024, 4, 1474331. https://doi.org/10.3389/fddsv.2024.1474331. Distributed under Open Access license CC BY.
  2. Kiessling, P.; Kuppe, C. Spatial multi-omics: novel tools to study the complexity of cardiovascular diseases. Genome Medicine 2024, 16, 14. https://doi.org/10.1186/s13073-024-01282-y. Distributed under Open Access license CC BY 4.0.
  3. Cao, J.; et al. Spatial transcriptomics: a powerful tool in disease understanding and drug discovery. 2024. PubMed Central: PMC11103497.
  4. Hao, Y.; Hao, S.; Andersen-Nissen, E.; et al. Integrated analysis of multimodal single-cell data. Cell 2021, 184, 3573–3587.e29. https://doi.org/10.1016/j.cell.2021.04.048.
  5. Luecken, M. D.; Theis, F. J. Current best practices in single-cell RNA-seq analysis: a tutorial. Molecular Systems Biology 2019, 15, e8746. https://doi.org/10.15252/msb.20188746.
  6. Squair, J. W.; Gautier, M.; Kathe, C.; et al. Confronting false discoveries in single-cell differential expression. Nature Communications 2021, 12, 5692. https://doi.org/10.1038/s41467-021-25960-2.
  7. Longo, S. K.; Guo, M. G.; Ji, A. L.; Khavari, P. A. Integrating single-cell and spatial transcriptomics to elucidate intercellular tissue dynamics. Nature Reviews Genetics 2021, 22, 627–644. https://doi.org/10.1038/s41576-021-00370-8.
  8. Moses, L.; Pachter, L. Museum of spatial transcriptomics. Nature Methods 2022, 19, 534–546. https://doi.org/10.1038/s41592-022-01409-2.

Online Inquiry

Submit your project details below, and our team will respond within 24 hours.

Molecular Dynamics Analysis
Free
Project Feasibility Assessment
Molecular Dynamics Analysis
Free
Project Sample
x
Need help getting the data you need?

Talk to our technical team about your project!

I Want To Talk