- Spatial multi-omics is an advanced approach in bioinformatics that combines multiple types of molecular information with their physical locations within tissues. It integrates spatially resolved measurements of different biological molecules, including DNA, RNA, proteins, metabolites, lipids, and epigenetic features, to provide a comprehensive understanding of tissue organization and cellular function. Unlike conventional molecular analysis, which often examines biological samples without preserving their original spatial arrangement, spatial multi-omics investigates both molecular composition and tissue location. This makes it possible to understand not only which biological processes occur but also where they occur and how they vary across different cellular populations and tissue regions.
- The concept of spatial multi-omics builds on two major developments: multi-omics and spatial biology. Multi-omics integrates complementary molecular layers to investigate biological systems at different levels, while spatial biology preserves information about the locations of cells and molecules within tissues. Combining these approaches allows researchers to connect genomic variation, gene regulation, RNA expression, protein abundance, metabolic activity, and tissue architecture. For example, a tumor tissue may contain genetically distinct cancer-cell populations, immune cells, fibroblasts, endothelial cells, and metabolically specialized regions. Spatial multi-omics can investigate how these components differ molecularly and how their locations influence interactions within the tumor microenvironment.
- Spatial transcriptomics is one of the most established components of spatial multi-omics. It measures RNA expression while retaining spatial coordinates or tissue context. Technologies such as 10x Visium, Visium HD, Slide-seq, MERFISH, seqFISH, Xenium, and CosMx provide different approaches and levels of spatial resolution. Some platforms measure gene expression across tissue spots, while others identify individual RNA molecules within cells or subcellular regions. Transcriptomic information can reveal cell types, gene-expression programs, cellular states, signaling pathways, and spatially variable genes. When integrated with other molecular layers, it helps connect RNA expression with protein abundance, regulatory activity, and metabolic function.
- Spatial genomics investigates DNA-level information within its tissue context. It can include the spatial distribution of genetic variants, somatic mutations, copy-number alterations, chromosomal abnormalities, and genomic heterogeneity. In cancer research, different tumor regions may contain genetically distinct cell populations because of clonal evolution and the accumulation of somatic mutations. Spatial genomic analysis can help investigate how these genetic differences are distributed across tumor regions and whether they are associated with particular transcriptional or protein-expression patterns. Techniques for spatially resolved DNA analysis are still developing, and the available resolution and molecular coverage depend strongly on the experimental platform.
- Spatial epigenomics examines regulatory features that influence gene activity without changing the underlying DNA sequence. These features include DNA methylation, histone modifications, chromatin accessibility, and other regulatory mechanisms. Methods based on spatially resolved chromatin profiling, imaging, or tissue-region-specific assays can help investigate how epigenetic states vary across anatomical regions and cellular environments. Spatial epigenomic information can be integrated with transcriptomic data to explore relationships between regulatory elements and gene expression. For example, differences in chromatin accessibility may help explain why cells in different tissue regions express distinct sets of genes despite sharing similar genomic sequences.
- Spatial proteomics measures proteins and their localization within tissues. Proteins are essential components of cellular structures, signaling pathways, enzymatic processes, immune responses, and tissue organization. Unlike RNA measurements, protein-level analysis provides information about the molecules that often directly perform biological functions. Technologies such as imaging mass cytometry, multiplexed ion beam imaging (MIBI), cyclic immunofluorescence, CODEX-based imaging, and other multiplexed antibody-imaging approaches can measure multiple proteins within tissue sections. Imaging-based platforms can identify cellular populations, protein localization, immune-cell organization, and signaling-associated patterns. Spatial proteomics is particularly useful for validating transcriptomic findings because RNA abundance does not always correspond directly to protein abundance.
- Spatial metabolomics investigates the distribution of small molecules involved in metabolism, energy production, cellular signaling, and biochemical reactions. Metabolites include amino acids, sugars, organic acids, nucleotides, lipids, and other metabolic intermediates. Their spatial distribution can reveal differences in metabolic activity between tissue regions or cellular environments. Mass spectrometry imaging methods, including matrix-assisted laser desorption/ionization mass spectrometry imaging (MALDI-MSI) and desorption electrospray ionization mass spectrometry imaging (DESI-MSI), can measure metabolites directly within tissue sections. These methods are useful for studying metabolic gradients, tumor metabolism, inflammation, tissue damage, and organ-specific biochemical functions.
- Spatial lipidomics focuses on the distribution and composition of lipids within tissues. Lipids contribute to cellular membranes, energy storage, cell signaling, and tissue structure. Spatial lipidomic analysis can identify differences in lipid composition between tissue compartments and investigate metabolic changes associated with disease, development, and cellular differentiation. Mass spectrometry imaging is an important approach for spatial lipidomics, although lipid identification, molecular coverage, and quantification can be challenging. Integrating lipidomic information with spatial transcriptomics and proteomics can help connect lipid metabolism with gene regulation and protein activity.
- Spatial multi-omics can also incorporate glycomics and other molecular layers. Glycans are carbohydrate structures attached to proteins and lipids that influence cell recognition, immune responses, adhesion, and molecular signaling. Spatial glycomic analysis can reveal how glycan patterns differ between tissue regions and cellular populations. Additional layers may include extracellular matrix composition, secreted factors, and molecular information about cell-surface structures. The selection of molecular layers depends on the biological question, available technologies, tissue type, and experimental resources.
- A major objective of spatial multi-omics is to integrate molecular measurements obtained from different technologies. These datasets may differ in spatial resolution, measurement sensitivity, gene or protein coverage, coordinate systems, and technical characteristics. Some methods measure several molecular layers simultaneously within the same tissue section, while others generate separate measurements from adjacent or serial tissue sections. Same-section measurements can provide stronger spatial correspondence, whereas serial-section integration requires image registration and careful alignment of anatomical structures. Differences in tissue processing and sample preparation can introduce additional technical variation.
- Spatial registration is an important step when integrating data from different tissue sections or imaging platforms. Registration aligns images or spatial coordinate systems so that corresponding anatomical structures and tissue regions can be compared. Image-processing tools such as Fiji/ImageJ, QuPath, CellProfiler, OpenCV, and SimpleITK can support image correction, segmentation, and registration. Accurate registration is essential because small spatial misalignments can lead to incorrect associations between molecular measurements from different modalities.
- Data preprocessing and quality control are also necessary before integration. Transcriptomic data may require sequencing quality assessment, normalization, filtering, and removal of technical artifacts. Protein-imaging data may require background correction, signal normalization, segmentation, and antibody-specific quality assessment. Mass spectrometry imaging may require spectral preprocessing, peak detection, normalization, and metabolite annotation. Differences in batch, tissue preparation, instrument performance, and detection sensitivity should be assessed separately for each molecular layer. Quality control is particularly important when combining measurements generated by different experimental technologies.
- Computational integration methods can be divided into several broad approaches. Early integration combines normalized molecular features from different omics layers into a shared analytical representation. Intermediate integration identifies common patterns or latent factors across molecular datasets. Late integration analyzes each data type separately and combines the resulting biological interpretations. Joint embedding methods represent samples, cells, or spatial locations in a shared low-dimensional space, while graph-based approaches connect cells or tissue regions according to molecular similarity and spatial proximity. Statistical models and machine-learning methods can be used to identify relationships among molecular layers and predict biological states.
- Several computational tools support multi-omics integration, although not all are specifically designed for spatial multi-omics. Seurat supports multimodal single-cell analysis and spatial transcriptomics workflows. Signac focuses on single-cell chromatin accessibility and can be integrated with other molecular modalities. MOFA+ identifies shared and modality-specific sources of variation across multiple omics datasets. MultiVI supports integration of single-cell gene-expression and chromatin-accessibility data. scvi-tools provides probabilistic and deep-learning frameworks for single-cell and multimodal data. Squidpy supports spatial graph analysis and spatial statistics, while Giotto provides spatial transcriptomics and spatial interaction analysis. SpatialExperiment provides structured data representation for spatially resolved experiments in R. The appropriate tool depends on the data types, spatial resolution, experimental design, and integration objective.
- Spatial multi-omics can also be combined with single-cell and single-nucleus sequencing. Single-cell RNA sequencing provides detailed cellular expression profiles, while single-cell ATAC sequencing measures chromatin accessibility. CITE-seq combines RNA expression with antibody-derived protein measurements, and other multimodal single-cell approaches can measure additional molecular features. These reference datasets can be mapped to spatial measurements to help identify cell types, cellular states, and regulatory programs within tissue sections. Deconvolution and reference-mapping methods such as cell2location, RCTD, Tangram, and related approaches can support the interpretation of spatial transcriptomic data, although their direct applicability depends on the spatial assay and molecular modalities being analyzed.
- An important application of spatial multi-omics is the study of cancer. Tumors are molecularly heterogeneous tissues containing malignant cells, immune populations, fibroblasts, endothelial cells, extracellular matrix, and other components. Different regions may contain distinct genetic alterations, transcriptional states, protein-expression patterns, and metabolic environments. Spatial multi-omics can investigate how these molecular differences are distributed across tumor regions and how they relate to invasion, metastasis, immune suppression, angiogenesis, and treatment resistance. Integrating genomic, transcriptomic, proteomic, and metabolomic measurements can provide a more complete understanding of tumor development and its surrounding microenvironment.
- In neuroscience, spatial multi-omics can help investigate molecular differences among brain regions, neuronal populations, glial cells, and specialized anatomical structures. Combining gene expression, protein localization, chromatin accessibility, and metabolite distribution can reveal molecular programs associated with neuronal function, neuroinflammation, development, and neurological disorders. In immunology, it can characterize the spatial organization of immune cells and their signaling environments during infection, inflammation, and immune responses. In developmental biology, it can reveal how gene regulation, protein expression, and metabolism vary across developing tissues and organs.
- Spatial multi-omics also has important applications in agriculture, veterinary medicine, and animal breeding. It can be used to investigate tissue organization and molecular mechanisms associated with animal growth, reproduction, immunity, metabolism, and adaptation. For example, combining transcriptomic and proteomic measurements in muscle tissue may help identify molecular pathways associated with muscle development and meat quality. Spatial analysis of reproductive tissues may reveal cell-specific molecular environments involved in fertility and development. In livestock disease research, integrated spatial measurements can help characterize host immune responses, pathogen-associated tissue changes, and cellular environments related to disease resistance. Comparative spatial multi-omics across breeds or species may also help identify conserved and species-specific biological mechanisms.
- Despite its potential, spatial multi-omics faces several technical and computational challenges. Different molecular technologies may have substantially different spatial resolutions, making direct comparison difficult. Some platforms measure multiple molecules simultaneously but have limited molecular coverage, while others provide detailed measurements of one molecular layer at a time. Tissue availability, sample preservation, assay costs, experimental complexity, and data storage requirements can also restrict large-scale studies. Integrating data from serial sections introduces uncertainty because the same cells may not be present in corresponding locations. In addition, different omics layers may have different detection limits and missing values.
- Another major challenge is computational complexity. Spatial multi-omics datasets can contain millions of molecular measurements across thousands of cells or tissue locations. Their analysis requires efficient data structures, statistical models, image-processing methods, and computational resources. Biological interpretation is also challenging because correlations between molecular layers do not necessarily demonstrate causal relationships. For example, a spatial association between a gene-expression pattern and a metabolite distribution may suggest a relationship but does not establish the underlying biochemical mechanism. Independent validation and functional experiments are therefore important.
- Overall, spatial multi-omics represents an important direction in bioinformatics because it combines molecular diversity with tissue architecture. By integrating genomic, epigenomic, transcriptomic, proteomic, metabolomic, and other molecular measurements within their spatial context, it provides a deeper understanding of cellular function, tissue organization, and biological interactions. Its integration with single-cell sequencing, advanced imaging, spatial transcriptomics, artificial intelligence, and computational modeling is expected to support increasingly detailed investigations of health, disease, development, veterinary biology, and animal genomics.