Tissue Architecture and Spatial Domains

  • Tissue architecture and spatial domains are important concepts in spatial transcriptomics and spatial omics because they describe how cells, molecular activities, and tissue structures are organized within biological tissues. Tissue architecture refers to the physical and functional arrangement of cells, extracellular matrix, blood vessels, nerves, and other structural components. Spatial domains are tissue regions characterized by distinct molecular profiles, cellular compositions, or biological functions. Together, these concepts help researchers understand how molecular activity is related to the organization of cells and anatomical structures.
  • Biological tissues are not simply collections of randomly distributed cells. Their cells are arranged in organized structures that support specific physiological functions. Epithelial tissues form protective and secretory layers, connective tissues provide structural support, muscle tissues generate force, and nervous tissues transmit information. Within organs, different cell populations occupy specific anatomical locations and interact with neighbouring cells, extracellular matrix, blood vessels, and signalling molecules. This organization is essential for normal tissue development, maintenance, repair, and function.
  • Tissue architecture is influenced by several biological components. Cellular composition determines which cell types are present within a tissue and their relative abundance. Cell morphology and orientation influence how cells are arranged and connected. The extracellular matrix provides structural support and biochemical signals that regulate cellular behaviour. Blood vessels deliver oxygen and nutrients, while nerves and immune cells contribute to tissue regulation and responses to environmental changes. Cell–cell adhesion, cell–matrix interactions, and local signalling pathways further establish and maintain tissue organization.
  • Spatial transcriptomics provides a way to investigate tissue architecture at the molecular level. Conventional gene-expression analysis generally measures RNA after it has been extracted from tissue, which removes information about the original locations of cells. Spatial transcriptomics combines gene-expression measurements with tissue coordinates, allowing researchers to determine where specific genes are expressed. Depending on the technology, measurements may represent spatial spots containing several cells, individual cells, or subcellular transcript locations. This information can be combined with histological images and other molecular measurements to characterize the organization of tissues.
  • Spatial domains are molecularly or biologically distinct regions within a tissue. They may correspond to recognizable anatomical structures, functional compartments, cellular neighbourhoods, or regions with specific physiological conditions. A spatial domain can be characterized by a particular combination of cell types, gene-expression patterns, molecular pathways, or tissue features. For example, a tumour may contain malignant-cell regions, stromal compartments, immune-cell-rich areas, and regions with limited oxygen availability. In the brain, spatial domains may correspond to anatomical structures and neuronal layers with distinct molecular characteristics.
  • Spatial domains can be identified using gene-expression profiles, spatial coordinates, histological information, cell-type composition, or combinations of these features. Computational clustering methods group spatial observations according to their molecular similarity and physical relationships. Graph-based methods construct neighbourhood networks in which cells or spots are represented as nodes and spatial relationships are represented as edges. Clustering algorithms such as Leiden and Louvain can then identify groups of related observations. Other approaches, including hierarchical clustering, k-means, and specialized spatial clustering models, can also be used depending on the dataset and research question.
  • Several computational methods have been developed specifically for spatial domain identification. BayesSpace uses Bayesian spatial modelling to identify spatial clusters and can improve the resolution of some spot-based spatial transcriptomics datasets. SpaGCN combines gene expression, spatial location, and histological information through graph convolutional networks. stLearn integrates spatial information and tissue morphology to support spatial clustering and analysis. BANKSY incorporates neighbourhood information into gene-expression representations to improve the identification of spatially organized cell populations. Seurat, Scanpy, Squidpy, and Giotto also provide tools for clustering, visualization, and spatial analysis.
  • Histological information is particularly valuable when defining tissue architecture and spatial domains. Hematoxylin and eosin (H&E) staining reveals general tissue morphology, including nuclei, cytoplasm, tissue boundaries, and structural organization. Immunohistochemistry and immunofluorescence can identify selected proteins and cell populations. These images provide anatomical context for interpreting spatial gene-expression patterns. Image-processing and analysis tools such as QuPath, Fiji/ImageJ, CellProfiler, and other computer-vision methods can support tissue annotation, image segmentation, and integration with spatial molecular measurements.
  • Spatial domains can also be characterized through spatially variable gene detection. Spatially variable genes are genes whose expression changes systematically according to their location within a tissue. Their expression may form gradients, localized regions, or patterns associated with specific anatomical structures. Methods such as SpatialDE, SPARK, and SPARK-X identify genes with statistically significant spatial expression patterns. Combining these results with clustering and histological annotation helps identify genes that characterize particular tissue compartments and reveal molecular differences between neighbouring regions.
  • Cell-type composition is another important feature of spatial domains. Different domains may contain different proportions of epithelial cells, fibroblasts, endothelial cells, immune cells, neurons, or other populations. In single-cell-resolution datasets, individual cells can be annotated using marker genes and reference-based methods. Tools such as SingleR, CellTypist, Seurat reference mapping, scmap, and Azimuth can support cell-type identification. In spot-based datasets, where each measurement may contain several cells, spatial deconvolution methods estimate the contributions of different cell populations. Important tools include RCTD, cell2location, SPOTlight, stereoscope, DestVI, and Tangram. These methods help researchers determine whether a spatial domain is associated with a particular cell population or a mixture of several cell types.
  • Tissue architecture is also influenced by cellular neighbourhoods and interactions. Cells communicate through secreted signalling molecules, direct cell contact, extracellular matrix interactions, and other mechanisms. Spatial analysis can identify which cell populations are located near one another and investigate whether particular cellular combinations occur more frequently in specific tissue regions. Ligand–receptor analysis tools such as CellChat, CellPhoneDB, NicheNet, and LIANA can help identify candidate signalling interactions between cell populations. Spatial proximity adds biological context to these predictions, although proximity and gene expression alone do not prove that a functional signalling interaction occurs.
  • An important aspect of tissue architecture is the organization of functional gradients. Gene expression and cellular activity may change gradually across a tissue rather than forming sharply separated regions. For example, the liver contains metabolic zones associated with differences in oxygen availability, nutrient supply, and metabolic activity. The brain contains molecular gradients across anatomical regions and neuronal layers. Developing tissues may exhibit gradients of signalling molecules that influence cell differentiation and tissue formation. Spatial transcriptomics can help identify these patterns by mapping gene expression across tissue coordinates and examining relationships between molecular activity and anatomical position.
  • Spatial domains can change during development, disease, tissue repair, and environmental adaptation. During embryonic development, cells become organized into distinct tissues and organs through coordinated gene regulation, cellular migration, proliferation, and differentiation. During wound healing, immune cells, fibroblasts, endothelial cells, and epithelial cells occupy different regions and contribute to tissue repair. In chronic inflammation, immune-cell populations may accumulate in specific tissue compartments. In cancer, tumour growth can alter normal tissue architecture and create spatially distinct regions with different cellular compositions, metabolic conditions, and molecular activities.
  • The analysis of tissue architecture and spatial domains often involves integrating several types of data. Spatial transcriptomics can be combined with single-cell RNA sequencing to improve cell-type identification and understand the molecular characteristics of spatially localized populations. Spatial proteomics provides information about protein abundance and localization, while histological imaging reveals tissue morphology. Spatial metabolomics can identify localized metabolic patterns. Multi-omics integration frameworks and software such as Seurat, Squidpy, Giotto, SpatialExperiment, and MOFA+ can support different aspects of these analyses. Integration must account for differences in measurement resolution, gene or protein coverage, sensitivity, and experimental design.
  • Tissue architecture and spatial domains have important applications in cancer research, neuroscience, developmental biology, immunology, infectious disease studies, and regenerative medicine. In cancer research, spatial analysis helps identify tumour compartments, invasive margins, immune-cell infiltration, stromal regions, and molecular patterns associated with treatment response. In neuroscience, it supports the mapping of neuronal and glial populations across brain structures. In developmental biology, it helps characterize the formation of tissue layers, organ structures, and cellular compartments. In immunology, it can reveal the organization of immune-cell populations within lymphoid and non-lymphoid tissues.
  • In animal genomics and breeding research, tissue architecture and spatial domains can be investigated in skeletal muscle, mammary glands, reproductive organs, liver, intestine, and immune tissues. Spatial analysis may reveal how different cell populations contribute to muscle development, muscle-fibre organization, meat quality, mammary gland function, milk production, reproductive performance, and disease resistance. Comparisons between breeds, developmental stages, physiological conditions, and environmental exposures may identify differences in tissue organization and molecular regulation associated with economically important traits. These findings can complement quantitative genetics and genomic studies by providing information about the cellular and tissue-level mechanisms underlying complex phenotypes.
  • Despite their value, the identification of spatial domains presents several challenges. Spatial clustering results depend on the measurement resolution, selected genes, normalization method, spatial neighbourhood definition, algorithm, and clustering parameters. Spot-based technologies may produce mixed expression profiles because several cells contribute to one spot. Single-cell-resolution technologies may be affected by incomplete transcript detection and inaccurate cell segmentation. Tissue processing can also cause deformation or loss of structural information. In addition, computationally identified domains do not always correspond to established anatomical structures, and some regions may contain continuous molecular gradients rather than clearly separated compartments.
  • Reliable interpretation requires integrating molecular measurements with tissue morphology, cell-type information, and biological knowledge. Researchers should inspect spatial maps, compare clustering results across suitable methods, examine marker genes, and assess the reproducibility of domains across biological replicates. Histological staining, immunofluorescence, immunohistochemistry, RNA in situ hybridization, and other experimental approaches can help validate the identity and function of spatially defined regions. It is also important to distinguish statistically identified spatial clusters from experimentally confirmed functional tissue compartments.
  • Overall, tissue architecture and spatial domains provide an essential framework for understanding the molecular and cellular organization of biological tissues. Spatial transcriptomics connects gene expression with physical tissue locations, while computational clustering, cell-type mapping, histological analysis, and spatial statistics help identify distinct tissue regions and their biological characteristics. Integrating these approaches with single-cell analysis, spatial proteomics, and other omics technologies provides a more complete understanding of tissue structure, cellular interactions, and biological function. These methods are increasingly important in biomedical research, veterinary science, and animal genomics because they connect molecular activity with the organization of complex biological systems.

Author:

Leave a Reply

Your email address will not be published. Required fields are marked *