Characterizing Tissue Structures from Spatial Omics with Spatial Cellular Graph Partition

Mammalian reproduction requires the fusion of two specialized cells: an oocyte and a sperm. In addition to producing gametes, the reproductive system also provides the environment for the appropriate development of the embryo. Deciphering the reproductive system requires understanding the functions of each cell type and cell-cell interactions. Recent single-cell omics technologies have provided insights into the gene regulatory network in discrete cellular populations of both the male and female reproductive systems. However, these approaches cannot examine how the cellular states of the gametes or embryos are regulated through their interactions with neighboring somatic cells in the native tissue environment owing to tissue disassociations. Emerging spatial omics technologies address this challenge by preserving the spatial context of the cells to be profiled. These technologies hold the potential to revolutionize our understanding of mammalian reproduction. We aim to review the state-of-the-art spatial transcriptomics (ST) technologies with a focus on highlighting the novel biological insights that they have helped to reveal about the mammalian reproductive systems in the context of gametogenesis, embryogenesis, and reproductive pathologies. We also aim to discuss the current challenges of applying ST technologies in reproductive research and provide a sneak peek at what the field of spatial omics can offer for the reproduction community in the years to come. The PubMed database was used in the search for peer-reviewed research articles and reviews using combinations of the following terms: 'spatial omics', 'fertility', 'reproduction', 'gametogenesis', 'embryogenesis', 'reproductive cancer', 'spatial transcriptomics', 'spermatogenesis', 'ovary', 'uterus', 'cervix', 'testis', and other keywords related to the subject area. All relevant publications until April 2023 were critically evaluated and discussed. First, an overview of the ST technologies that have been applied to studying the reproductive systems was provided. The basic design principles and the advantages and limitations of these technologies were discussed and tabulated to serve as a guide for researchers to choose the best-suited technologies for their own research. Second, novel biological insights into mammalian reproduction, especially human reproduction revealed by ST analyses, were comprehensively reviewed. Three major themes were discussed. The first theme focuses on genes with non-random spatial expression patterns with specialized functions in multiple reproductive systems; The second theme centers around functionally interacting cell types which are often found to be spatially clustered in the reproductive tissues; and the thrid theme discusses pathological states in reproductive systems which are often associated with unique cellular microenvironments. Finally, current experimental and computational challenges of applying ST technologies to studying mammalian reproduction were highlighted, and potential solutions to tackle these challenges were provided. Future directions in the development of spatial omics technologies and how they will benefit the field of human reproduction were discussed, including the capture of cellular and tissue dynamics, multi-modal molecular profiling, and spatial characterization of gene perturbations. Like single-cell technologies, spatial omics technologies hold tremendous potential for providing significant and novel insights into mammalian reproduction. Our review summarizes these novel biological insights that ST technologies have provided while shedding light on what is yet to come. Our review provides reproductive biologists and clinicians with a much-needed update on the state of art of ST technologies. It may also facilitate the adoption of cutting-edge spatial technologies in both basic and clinical reproductive research.

Spatial transcriptomics has emerged as a transformative technology in biomedical research, offering unprecedented insights into gene and protein expression within their native tissue context. Unlike conventional bulk or single-cell sequencing approaches, spatial omics has the advantage of preserving the spatial structure of tissues, allowing researchers to directly map molecular information onto histological structures. This review provides an overview of the current state of spatial omics technologies, highlighting their application in cancer research. Spatial omics has enabled detailed characterization of the tumor microenvironment (TME), revealing spatial heterogeneity, immune cell infiltration patterns, and complex mechanisms of tumor progression and therapy resistance across various cancer types. The review covers future directions, including artificial intelligence–driven analytics, improved standardization, and cost reduction to accelerate clinical translation. Ultimately, spatial omics is poised to play a central role in precision oncology, enabling a deeper understanding of tumor biology and informing more effective individualized treatment strategies.

Diabetic kidney disease (DKD), a primary cause of end-stage renal disease, results from progressive tissue remodeling and loss of kidney function. While single-cell RNA sequencing has significantly accelerated our understanding of cellular diversity and dynamics in DKD, its lack of spatial resolution limits insights into tissue-specific dysregulation and the microenvironment. Spatial transcriptomics (ST) is an innovative technology that combines gene expression with spatial localization, offering a powerful approach to dissect the molecular mechanisms of DKD. This mini-review introduces how ST has transformed DKD research by enabling spatially resolved analysis of cell interactions and identifying localized molecular alterations in glomeruli and tubules. ST has revealed dynamic intercellular communication within the renal microenvironment, lesion-specific gene expression patterns, and immune infiltration profiles. For example, Slide-seqV2 has highlighted disease-specific cellular neighborhoods and associated signaling networks. Furthermore, ST has pinpointed key genes implicated in disease progression, such as fibrosis-related proteins and transcription factors in tubular damage. By integration of ST with computational tools such as machine learning and network-based analysis can help uncover gene regulatory mechanisms and potential therapeutic targets. However, challenges remain in limited spatial resolution, high data complexity, and computational demands. Addressing these limitations is essential for advancing precision medicine in DKD.

The identification of key pathogenic cytokines and association of certain immune pathways with distinct inflammatory skin diseases have paved the way for targeted therapies and novel diagnostic tools.1 While bulk transcriptomic analyses represent a cost-effective and accessible method to measure average gene expression within a lesion, this approach obscures key layers of information: cell type heterogeneity and potentially crucial spatial organization. Inflammation is not evenly distributed; rather, it is tightly compartmentalized within skin lesions. These microenvironmental niches can influence which cells interact, which structures are involved, and how accessible they are to therapeutic intervention. Jiang and colleagues contend that lesions should be regarded as collections of inflammatory compartments rather than as uniformly inflamed tissue.2 A niche constitutes more than a mere collection of cells. It reflects a spatially organized circuit, composed of immune cells, vessels, stroma and epithelium, that contributes to disease memory and recurrence and enables chronic disease progression. Using hidradenitis suppurativa and psoriasis as an example, the authors elucidate how spatial transcriptomics and the concept of inflammatory niches help to understand why these two diseases—despite sharing common key inflammatory pathways and pathogenic cytokines—display different response rates to targeted therapies.2 The authors present a persuasive argument that these spatial patterns function as determinants of skin disease. Indeed, this concept is already reshaping our understanding of other inflammatory conditions. Recent spatial data in primary cicatricial alopecia reveal a hostile, follicle-centric environment characterized by immune cells actively compromising epithelial immune privilege.3 These distinctions are not solely superficial; they represent discrete microanatomical circuits. The authors also emphasize the presence of compartmentalized cytotoxic circuits at the dermo-epidermal junction in vitiligo, along with growing neuro-immune influences that modulate dendritic cell function and local CD8+ T cell activity.2 With the identification of Cathepsin S-expressing melanocytes situated within specific compartments alongside antigen-presenting cells,4 this finding transforms the inflammatory niche from a promising discovery into a viable therapeutic target: to identify and interrupt circuits within defined tissue regions. The concept of the inflammatory niche raises the bar for personalized medicine, and spatial transcriptomics provides unprecedented insight into tissue-level heterogeneity in inflammatory skin diseases. By integrating gene expression with tissue localization, spatial transcriptomics enables direct correlation between molecular profile and histopathologic structure, allowing us to accurately map the structural organization of tissue inflammation. However, a note of caution is still warranted. Spatial omics has limitations: low-abundance transcripts may be overlooked, cell identities often rely on marker gene inference, and cost and technical demands remain barriers to broader implementation.5 It will be interesting to see if artificial intelligence-powered analyses can soon reverse the current paradigm and infer specific immune pathways and inflammatory niches from regular histopathology alone, thereby bringing precision medicine closer to everyday clinical routine. Ultimately, whether treatment selection based on niche-specific targets can improve outcomes—and whether the concept of inflammatory niches will evolve into a practical tool for therapeutic decision-making—remains an open and important question. The author(s) received no financial support for the research, authorship and/or publication of this article. JLN: Speaker honoraria—Novartis and Sun Pharmaceuticals. CC: Scientific adviser and/or clinical study investigator—AbbVie, Actelion, Almirall, Amgen, Boehringer Ingelheim, Bristol Myers Squibb, Celgene, Eli Lilly, Galderma, Incyte, Jiangsu Hengrui, Johnson & Johnson, LEO Pharma, MSD, Novartis, Pfizer, Samsung, Sanofi, Takeda and UCB. Not applicable. Data sharing is not applicable to this article as no new data were created or analyzed in this study.

Diabetic nephropathy, a pathologically diagnosed microvascular complication of diabetes, is a strong risk factor for cardiovascular events, which mainly involve arteries larger than those affected in diabetic nephropathy. However, the association between diabetic nephropathy pathological findings and cardiovascular events has not been well studied. We aimed to investigate whether the pathological findings in diabetic nephropathy are closely associated with cardiovascular event development. This retrospective cohort study analysed 377 people with type 2 diabetes and biopsy-proven diabetic nephropathy, with a median follow-up of 5.9 years (interquartile range 2.0 to 13.5). We investigated how cardiovascular events were impacted by two vascular diabetic nephropathy lesions, namely arteriolar hyalinosis and arterial intimal thickening, and by glomerular and interstitial lesions. Of the 377 people with diabetic nephropathy, 331 (88%) and 295 (78%) had arteriolar hyalinosis and arterial intimal thickening, respectively. During the entire follow-up period, those with arteriolar hyalinosis had higher cardiovascular event rates in the crude Kaplan-Meier analysis than those without these lesions (P=0.005, log-rank test). When fully adjusted for clinically relevant confounders, arteriolar hyalinosis independently predicted cardiovascular events [hazard ratio (HR) 1.99; 95% confidence interval (CI) 1.12, 3.86], but we did not find any relationship between arterial intimal thickening and cardiovascular events (HR 0.89; 95% CI 0.60, 1.37). Additionally, neither glomerular nor interstitial lesions were independently associated with cardiovascular events in the fully adjusted model. Arteriolar hyalinosis, but not intimal thickening of large arteries, was strongly associated with cardiovascular events in people with diabetic nephropathy.

Spatial omics (SOs) are powerful methodologies that enable the study of genes, proteins, and other molecular features within the spatial context of tissue architecture. With the growing availability of SO data sets, researchers are eager to extract biological insights from larger data sets for a more comprehensive understanding. However, existing approaches focus on batch effect correction, often neglecting complex biological patterns in tissue slices, complicating feature integration and posing challenges when combining transcriptomics with other omics layers. Here, we introduce spatial multislice/omics analysis (stMSA), a deep graph contrastive learning model that incorporates graph auto-encoder techniques. stMSA is specifically designed to produce batch-corrected representations while retaining the distinct spatial patterns within each slice, considering both intra- and inter-batch relationships during integration. Extensive evaluations show that stMSA outperforms state-of-the-art methods in distinguishing tissue structures across diverse slices, even when faced with varying experimental protocols and sequencing technologies. Furthermore, stMSA effectively deciphers complex developmental trajectories by integrating spatial proteomics and transcriptomics data and excels in cross-slice matching and alignment for 3D tissue reconstruction.

Rapid advances in spatial transcriptomics (ST) have revolutionized the interrogation of spatial heterogeneity and increase the demand for comprehensive methods to effectively characterize spatial domains. As a prerequisite for ST data analysis, spatial domain characterization is a crucial step for downstream analyses and biological implications. Here we propose a prior-based self-attention framework for spatial transcriptomics (PAST), a variational graph convolutional autoencoder for ST, which effectively integrates prior information via a Bayesian neural network, captures spatial patterns via a self-attention mechanism, and enables scalable application via a ripple walk sampler strategy. Through comprehensive experiments on data sets generated by different technologies, we show that PAST can effectively characterize spatial domains and facilitate various downstream analyses, including ST visualization, spatial trajectory inference and pseudotime analysis. Also, we highlight the advantages of PAST for multislice joint embedding and automatic annotation of spatial domains in newly sequenced ST data. Compared with existing methods, PAST is the first ST method that integrates reference data to analyze ST data. We anticipate that PAST will open up new avenues for researchers to decipher ST data with customized reference data, which expands the applicability of ST technology.

Spatial omics technologies offer unprecedented insights into the cellular organization of tissues; however, they are not yet scalable for routine clinical use. In contrast, images of histological staining remain the foundation of pathological diagnosis despite lacking molecular information. Bridging this gap requires computational methods that can accurately infer spatial molecular data from histology alone. Here, we introduce TissueCraftAI, a generative artificial intelligence framework that predicts multi-modal spatial omics maps directly from standard histology images using natural language prompts. To train and validate our model, we created PRISM-12M, a large-scale dataset comprising over twelve million spatially registered histology and spatial omics image patches across fourteen tissue types from humans and mice. TissueCraftAI significantly outperforms existing methods in generating realistic histology images and predicting spatial proteomics and transcriptomics data with high fidelity. We demonstrated its utility in various downstream applications, including improving cell type annotation and enhancing the accuracy of patient survival predictions across multiple cancer types. By enabling flexible, query-driven in silico spatial molecular analysis using routine histology images, TissueCraftAI opens up new research avenues in computational pathology.

Since its first application in 2016, spatial transcriptomics has become a rapidly evolving technology in recent years. Spatial transcriptomics enables transcriptomic data to be acquired from intact tissue sections and provides spatial distribution information and remedies the disadvantage of single-cell RNA sequencing (scRNA-seq), whose data lack spatially resolved information. Presently, spatial transcriptomics has been widely applied to various tissue types, especially for the study of tumor heterogeneity. In this review, we provide a summary of the research progress in utilizing spatial transcriptomics to investigate tumor heterogeneity and the microenvironment with a focus on solid tumors. We summarize the research breakthroughs in various fields and perspectives due to the application of spatial transcriptomics, including cell clustering and interaction, cellular metabolism, gene expression, immune cell programs and combination with other techniques. As a combination of multiple transcriptomics, single-cell multiomics shows its superiority and validity in single-cell analysis. We also discuss the application prospect of single-cell multiomics, and we believe that with the progress of data integration from various transcriptomics, a multilayered subcellular landscape will be revealed.

Spatial transcriptomic (ST) technologies, such as GeoMx Digital Spatial Profiler, are increasingly utilized to investigate the role of diverse tumor microenvironment components, particularly in relation to cancer progression, treatment response, and therapeutic resistance. However, in many ST studies, the spatial information obtained from immunofluorescence imaging is primarily used for identifying regions of interest (ROIs) rather than as an integral part of downstream transcriptomic data analysis and interpretation. We developed ROICellTrack, a deep learning-based framework that better integrates cellular imaging with spatial transcriptomic profiling. By analyzing 56 ROIs from urothelial carcinoma of the bladder and upper tract urothelial carcinoma, ROICellTrack identified distinct cancer-immune cell mixtures, characterized by specific transcriptomic and morphological signatures and receptor-ligand interactions linked to tumor content and immune infiltrations. Our findings demonstrate the value of integrating imaging with transcriptomics to analyze spatial omics data, improving our understanding of tumor heterogeneity and its relevance to personalized and targeted therapies. ROICellTrack is publicly available at https://github.com/wanglab1/ROICellTrack.

Graph Convolutional Networks (GCNs) are widely applied for spatial domain identification in spatial transcriptomics (ST), where node representations are learned by aggregating information from neighboring spots. However, most ST workflows construct spatial graphs by assigning equal weights to neighbors and self-loops, and then applying degree-based normalization. This procedure often yields near-uniform adjacency matrices, suppressing natural distance heterogeneity, diminishing spatial resolution, aggravating GCN over-smoothing, and obscuring fine-grained tissue boundaries. We introduce DWGCN, a Distance-Weighted Graph Convolutional Network that replaces uniform neighbor assignment with inverse-distance weighting (IDW) and spot-wise normalization. DWGCN enhances locality-sensitive aggregation by assigning larger weights to proximal neighbors, while preserving self-loop dominance to maintain intrinsic spot information and reduce hub-driven dilution. Across four real and four simulated ST datasets, integrating DWGCN with representative GCN-based frameworks (SEDR, GraphST, SpaNCMG, SpaGIC) generally improved clustering accuracy, particularly in tissues with complex spatial architectures. These results demonstrate that DWGCN offers a broadly applicable approach for restoring distance-aware structure in spatial graphs, thereby improving the delineation of spatial domain identification.

Recent advances in self-supervised foundation models for computational cancer histopathology have shown great promise for predicting clinical phenotypes from routine hematoxylin and eosin (H&E)-stained images. However, application of these models in the context of discovery, for example for identification of novel biomarkers or drug targets, remains challenging as this typically requires linking whole slide imaging (WSI) data to molecular data, which has historically been scarce. Recently, spatial omics (SO) has emerged as a powerful technology for high-throughput profiling of RNAs and proteins directly within cancer tissues. This presents new and exciting opportunities to improve histopathology models for making molecular predictions about the tumor microenvironment. In this work, we develop machine learning models that can generate transcriptomic profiles at two different spatial scales: (1) cellular-level resolution, i.e. SO, and (2) slide-level resolution, i.e. bulk RNA-sequencing. We present a preprocessing and benchmarking framework to enable rigorous comparison of models on an in-house SO dataset, as well as a public bulk RNA-seq dataset (TCGA-COAD/READ), of colorectal cancer. We fine-tune recently published digital pathology foundation models for these tasks, and find that they outperform both general pathology-naïve image foundation models, as well as specialized digital pathology deep learning models. We also find that existing models cannot be simultaneously trained using supervision from multiple spatial resolutions. To address this, we propose a graph neural network (GNN) modeling approach that can model both local interactions between cells within tumor niches, as well as global interactions across tumor regions. We show that our GNN model outperforms existing models for both cell-level and slide-level tasks. This work presents a novel multi-scale framework for predicting transcriptome profiles from WSIs. Citation Format: Tianyi Zhang, Malay Singh, Dwivedi Naman, Timothy Antoni, Ju Can, Firas Elbayoumi, Iain Bee Huat Tan, Tiantian He, Hwee Kuan Lee, Shyam Prabhakar, Grace Hui Ting Yeo. Multi-Scale Modeling for Prediction of Spatial Single-Cell and Bulk Transcriptomic Profiles from Whole-Slide Images of Colorectal Cancer [abstract]. In: Proceedings of Frontiers in Cancer Science 2024; 2024 Nov 13-15; Singapore. Philadelphia (PA): AACR; Cancer Res 2025;85(15_Suppl):Abstract nr P24.

e15076 Background: Precision oncology aims to first break diagnoses into biologically distinct subtypes, then pursue personalized therapies for each group of patients ( Garraway et al, J Clin Onc, 2013). This strategy is enabled by recent advances in technologies for medical imaging, molecular profiling, and artificial intelligence. Spatially resolved molecular profiling (“spatial omics”) is an emerging technology that harnesses all three of these advances ( Navarro et al, Science, 2016; Rao et al, Nature, 2021). Cancer is driven by localized interactions between tumor cells, immune and non-immune stromal components. Methods: Spatial omics studies have revealed specific cell types and pathways that control those tumor-microenvironment crosstalks ( Hunter et al, Nat Comm, 2021; Alon et al, Science, 2021), involving both innate and adaptive immunity ( Binnewies et al, Nat Med, 2018). Identifying new cancer subtypes based on the spatially localized tumor-host interactions across many patients have the potential to revolutionize immuno-oncology. The advances in utilizing AI-based methods have so far been unattainable due to a small number of patients in spatial omics studies. Results: Here we present the MOSAIC project - a collaborative initiative across industry and top oncology hospitals to build the largest collection of spatial omics data in cancer. We combine comprehensive clinical annotation with new deep profiling methods to both discover cancer subtypes and identify drug targets and biomarkers within them. MOSAIC aims to generate multimodal data for a total of 7,000 patients in seven cancer indications. Data modalities will include spatial and single cell transcriptomics and proteomics, bulk molecular profiling, pathology images, and curated clinical information. This collaboration harnesses the strengths of academia and industry to provide patient samples, generate high-quality data, develop AI-based analytical tools, and compile a resource that will eventually be made widely available for medical research. Conclusions: Here we present the workflow of the MOSAIC study, demonstrate the data processing pipelines and show spatial distributions of transcripts within the tumor microenvironment. Building upon these initial datasets, we discuss how the MOSAIC parties will collaborate to build an unprecedented database for research and discovery of new immuno-oncology therapeutic approaches.

Spatially resolved transcriptomics involves a set of emerging technologies that enable the transcriptomic profiling of tissues with the physical location of expressions. Although a variety of methods have been developed for data integration, most of them are for single-cell RNA-seq datasets without consideration of spatial information. Thus, methods that can integrate spatial transcriptomics data from multiple tissue slides, possibly from multiple individuals, are needed. Here, we present PRECAST, a data integration method for multiple spatial transcriptomics datasets with complex batch effects and/or biological effects between slides. PRECAST unifies spatial factor analysis simultaneously with spatial clustering and embedding alignment, while requiring only partially shared cell/domain clusters across datasets. Using both simulated and four real datasets, we show improved cell/domain detection with outstanding visualization, and the estimated aligned embeddings and cell/domain labels facilitate many downstream analyses. We demonstrate that PRECAST is computationally scalable and applicable to spatial transcriptomics datasets from different platforms.

Spatially mapping the transcriptome and proteome in the same tissue section can profoundly advance our understanding of cellular heterogeneity and function. Here we present Deterministic Barcoding in Tissue sequencing plus (DBiTplus), an integrative multimodal spatial omics approach combining sequencing-based spatial transcriptomics and multiplexed protein imaging on the same section, enabling both single-cell-resolution cell typing and transcriptome-wide interrogation of biological pathways. DBiTplus utilizes spatial barcoding and RNase H-mediated cDNA retrieval, preserving tissue architecture for multiplexed protein imaging. We developed computational pipelines to integrate these modalities, allowing imaging-guided deconvolution to generate single-cell-resolved spatial transcriptome atlases. We demonstrate DBiTplus across diverse samples including frozen mouse embryos, and formalin-fixed paraffin-embedded human lymph nodes and lymphoma tissues, highlighting its compatibility with challenging clinical specimens. DBiTplus uncovered mechanisms of lymphomagenesis, progression and transformation in human lymphomas. Thus, DBiTplus is a unified workflow for spatially resolved single-cell atlasing and unbiased exploration of biological mechanisms in a cell-by-cell manner at transcriptome scale.