Why Spatial Multiomics? What RNA and Protein Measurements Add

Measuring RNA and protein expression together within the same tissue section enables researchers to interrogate the most appropriate analyte according to the question. This can improve the identification of cell populations, characterisation of activity states, and investigation into how transcriptional profiles relate to protein expression and cell and tissue function.  

Preserving spatial information adds another dimension to connect these measurements to biological questions which depend on tissue architecture. In immune engager or antibody-drug conjugate studies, the localisation of proteins to cell surfaces can be examined, or in cell-cell communication studies, the proximity of receptor and ligand on neighbouring cells can be correlated to downstream transcriptional activation.

For biopharma teams, in situ imaging technologies such as Xenium, CosMx and COMET provide the tools to answer a diverse range of questions from a single sample. Propath facilitates studies on these platforms, from sample preparation to data analysis, with expert guidance on assay selection and study design as well as support for technology assessment, biomarker discovery, validation and mechanism-of-action studies.

RNA and protein provide complementary evidence

mRNA provides a highly scalable analyte for interrogating cell lineage, activity and transcriptional state, from focused gene signatures to whole-transcriptome, single cell capability. Proteins are an order of magnitude more complex, due to a more complex alphabet and 3D-folded conformations. The generation and validation of detection probes (typically antibody-based) is similarly more complex and less suitable for larger studies.  

At a technical level, spatial transcriptomics provides a highly scalable solution, while the inclusion of targeted antibody panels provides key insights into effector-level changes, protein-protein interactions, post-translational localisation and degradation, and extracellular aggregates.

Beyond technical considerations, measuring either protein or RNA alone can limit the conclusions that can be drawn from a study. Regulation between transcription and protein production together with differences in molecular turnover mean that transcript measurement does not necessarily correlate with protein expression levels.

Likewise, detecting a protein does not establish that it is active, accessible to a therapeutic agent or driving a downstream response. While these questions may require targeted functional assays such as proximity ligation assays, the additional dimension of information provided by transcriptional states and spatial context can offer insight.  

Transcriptomics also offers information on the production of secretory cytokines and chemokines, as well as transient or labile proteins that may not survive sample preparation or non-coding regulatory elements such as siRNA.  

Cell phenotyping is augmented by multiomic analysis  

Classical histological cell typing using single or multiple protein expression markers can struggle to resolve some functional phenotypes such as Schwann cells which are not specifically typed by either S100 or SOX10 without co-detection of MPZ or PMP22 mRNA expression.  

Cell phenotyping by transcriptomic signatures is less sensitive to the absence of one key lineage-defining protein, but section thickness and difficulties accurately segmenting cells in situ can leave some ambiguity at the periphery of transcriptional clusters.  

Combining the strengths of RNA and protein for a multi-modal analysis approach enables one layer to augment and often supervise the other, providing coverage where the other is less well suited.  

Tissue location changes the question researchers can ask

In situ imaging measures molecular targets within a tissue section while retaining their spatial coordinates. This allows expression data to be examined alongside tissue morphology, cell boundaries and local tissue organisation. Foundational molecular imaging studies have demonstrated that RNA and protein imaging at subcellular resolution in fixed tissue can provide deeper insights than molecular signatures observed in the absence of spatial context.

Consider a tumour study in which a candidate target’s RNA and protein are both detected. A bulk analysis result leaves several questions open. Are the signals associated with malignant cells, infiltrating immune cells, stroma or adjacent tissue? Are positive cells distributed throughout the lesion, organised in structures such as an angiogenic vessel or an emerging lymphoid aggregate or restricted at its edge? Do the RNA and protein signals occur in the same cells? Can other differential signals seen be correlated to the same cells, or are they co-incidental?

These are distinct research questions. The first concerns the source of expression, the second concerns heterogeneity, the third concerns agreement between modalities and the fourth concerns mechanism. Leaving the tissue intact and imaging the analytes in situ keeps these hypotheses testable.

When developing a research framework, consider what each layer of information contributes:

Selecting a Technology Platform

Xenium, CosMx, and COMET all support imaging-based spatial research, including workflows that combine RNA and protein measurements. When comparing platforms, teams should consider the underlying chemistry, gene and protein coverage requirements, and downstream analysis workflows.  

10x Xenium

By employing a padlock probe design, Xenium RNA probes provide dual-recognition capable of high specificity and robust performance in low-quality RNA. Hybridised probes are detected via ligation and rolling-circle amplification, followed by multiple rounds of fluorescent imaging and decoding. Gene coverage ranges from a suite of focused, customisable panels capable of co-detecting up to 480 RNA targets, to the Xenium Prime 5K panel, supporting targeted investigations and broader transcriptional profiling in both human and mouse.

For combined RNA–protein co-detection, compatible human panels targeting fewer than 480 RNA targets can be paired with up to 27 protein targets in the same FFPE section. Downstream bioinformatic analysis can connect transcriptional profiles with protein translation and localisation. While current workflows support multimodal staining and deep-learning segmentation, with the protein layer supporting refined cell boundary detection for improved transcript assignment, these methods are highly tissue specific and require review and refinement before interpreting biological patterns.

Bruker Spatial CosMx

CosMx also profiles RNA through probe hybridisation and repeated fluorescent barcode imaging. While the more traditional probes’ higher affinity provides robust sensitivity in poor-quality RNA found in FFPE tissue, this higher-affinity hybridisation means high-homology variants can be difficult to distinguish. Direct fluorescent labelling of these probes simplifies the assay by avoiding the requirement for enzymatic amplification and PCR cycling and any associated bias, and unlocks up to whole-transcriptome scale.

Human and mouse panels range from targeted, customisable 1,000-gene panels to whole-transcriptome assays in both human and mouse formats, covering more than 19,000 RNA targets and enabling broad investigation of cell identities and transcriptional programmes within their tissue context. However, it’s important to note that panel coverage does not guarantee detection of every target in every cell – 5µm sampling of cells and optical crowding potential at high panel depths can impact per-gene-per-cell sensitivity compared to more targeted approaches.

Compatible human RNA assays can be combined with up to 72 protein targets in the same FFPE section, linking transcriptional information with selected protein phenotypes. Downstream analysis uses z-stacked imaging of key morphology markers and machine-learning segmentation to provide 3D cell segmentation through the AtoMx cloud platform, reducing the miss-assignment of transcripts to overlapping cells in a 5µm section using non-vertical z-plane boundaries.  

Lunaphore COMET

Integrated microfluidics and fluorescent slide scanning enable fully custom antibody and RNA panels by immunoflouresecense and in situ hybridization (ISH). For more targeted studies focused on custom panels, co-registered ISH can provide RNA expression data for key targets. Propath has experience in optimisation and validation of signature multiomics assays of up to 12 RNA targets plus up to 24 protein targets to support applications such as tracing viral vectors or CAR-T cells and investigating cytokine expression.

More focused data generation, classical IHC/ISH methodology, higher sample throughput and lower cost per sample all make cyclic immunoflourescence a robust and translatable solution where assessment of  a broad ceullar response or a targeted pathway deep-dive is indicated.

Xenium and CosMx are well suited to studies that prioritise broad, customisable RNA profiling with complementary protein measurements that are more limited. COMET offers an alternative for studies driven by custom protein panels, with targeted RNA detection to address specific questions. At Propath, same-section multiplex fluorescence staining provides a fully customisable proteomics data layer, including proximity ligation assays for interactomics readouts. The most appropriate approach depends on the RNA coverage, protein targets and spatial relationships needed to answer the study question.

Design the study around the interpretation you need

Combining modalities is most useful when each measurement has a defined purpose. Before selecting in situ imaging or multiplex immunofluorescence methods, identify the cell populations, transcriptional states, protein phenotypes, tissue relationships, per-target sensitivity and sample number needed to answer the question.

Then decide whether same-cell co-localisation is essential, whether a predeveloped panel provides adequate biological coverage or whether a custom panel is required, and which comparisons will distinguish biological variation from technical effects. Biological replication remains important: large numbers of cells from one specimen do not replace independent samples, and studies should be powered according to the effect size expected.

The benefit of spatial multiomics comes from the relationships it allows researchers to examine: between RNA and protein, between neighbouring cells, between tissue contexts, between normal and disease and between treated and untreated – all from a single assay. Defining those relationships early helps determine which measurements are worth making and what evidence will be needed to interpret them. Don’t leave bioinformatic planning until after data is generated, as over-generating data can unnecessarily overload data pipelines and delay results.

Planning a spatial multiomics study?

Propath offers multiomics services using Xenium, CosMx and COMET alongside platform-agnostic spatial bioinformatics, including custom cell-segmentation, cell typing, neighbourhood analysis, differential pathway analysis, pseudotime and integration of spatial and non-spatial data. These capabilities support study planning that connects assay selection with the intended analysis.

Discuss your biological question with Propath to explore whether combining spatial RNA and protein measurements could strengthen your study, and which assays fit your tissue and research objectives.