Evidence-aware Integration and Domain Identification of Spatial Transcriptomics Data

Spatial transcriptomics (ST) technologies enable the measurement of mRNA expression while simultaneously capturing spot locations. By integrating ST data, the 3D structure of a tissue can be reconstructed, yielding a comprehensive understanding of the tissue's intricacies. Nevertheless, a computational challenge persists: how to remove batch effects while preserving genuine biological structure variations across ST data. To address this, we introduce Graspot, a graph attention network designed for spatial transcriptomics data integration with unbalanced optimal transport. Graspot adeptly harnesses both gene expression and spatial information to align common structures across multiple ST datasets. It embeds multiple ST datasets into a unified latent space, facilitating the partial alignment of spots from different slices. Demonstrating superior performance compared to existing methods on four real ST datasets, Graspot excels in ST data integration, including tasks that require partial alignment. In particular, Graspot efficiently integrates multiple ST slices and guides coordinate alignment. In addition, Graspot accurately aligns the spatio-temporal transcriptomics data to reconstruct human heart developmental processes. Graspot software is available at https://github.com/zhan009/Graspot.

Spatial transcriptomics (ST) is a cutting-edge technology that provides comprehensive insights into gene expression patterns from a spatial perspective. A key research focus within this field is spatial domain identification, which is essential for exploring tissue organization, biological development, and disease mechanisms. Although methods have been developed, they still face challenges in modeling the gene expression information together with the spatial locations, resulting in suboptimal accuracy. We introduce Distribution-aware Contrastive Learning for Spatial Transcriptomics (DisConST), a novel deep learning method designed to improve spatial domain detection within ST datasets. DisConST addresses key challenges, such as the high dropout rates and the complex integration of spatial and gene expression data, by incorporating contrastive learning strategies that are aware of the underlying data distributions. It employs the zero-inflated negative binomial (ZINB) distribution, along with graph contrastive learning, to generate more informative latent representations. These representations efficiently integrate spatial positions, transcriptomic profiles, and cell-type proportions within spots. We validated DisConST across diverse ST datasets of tissues, organs, and embryos from various sequencing platforms in both normal and disease states. Our results consistently demonstrated that DisConST achieves superior spatial domain recognition accuracy compared to existing state-of-the-art methods. Furthermore, our experiments highlighted the utility of DisConST in advancing research on tissue organization, embryonic development, and tumor immune microenvironment dissection. The source code for DisConST is freely available at https://github.com/Zhenpm/DisConST/.

Spatial transcriptomics is a rapidly developing field of single-cell genomics that quantitatively measures gene expression while providing spatial information within tissues. A key challenge in spatial transcriptomics is identifying spatially structured domains, which involves analyzing transcriptomic data to find clusters of cells with similar expression patterns and their spatial distribution. To address these challenges, we propose a novel deep-learning method called DMGCN for domain identification. The process begins with preprocessing that constructs two types of graphs: a spatial graph based on Euclidean distance and a feature graph based on Cosine distance. These graphs represent spatial positions and gene expressions, respectively. The embeddings of both graphs are generated using a multi-view graph convolutional encoder with an attention mechanism, enabling separate and co-convolution of the graphs, as well as corrupted feature convolution for contrastive learning. Finally, a fully connected network (FCN) decoder is employed to generate domain labels and reconstruct gene expressions for downstream analysis. Experimental results demonstrate that DMGCN consistently outperforms state-of-the-art methods in various tasks, including spatial clustering, trajectory inference, and gene expression broadcasting.

Background: Spatial transcriptomics (ST) enables the integration of gene expression data with spatial context in tissue samples, providing high-resolution insights without requiring single-cell dissociation. With subcellular resolution, ST offers a precise method for annotating cell types, surpassing traditional approaches based on manual pathologic annotation. Conventional AI models for pathology images typically rely on annotations from pathologists to label cellular compositions in hematoxylin and eosin (H&E)-stained slides. In contrast, our AI model leverages training data generated from subcellular resolution ST, providing a distinct advantage by achieving higher accuracy and objectivity in identifying specific cell types. In this study, we focus on lymphocyte identification in non-small cell lung cancer (NSCLC) pathology images to demonstrate the capabilities of this approach. Methods: This model analyzes H&E slide images obtained from surgical specimens of NSCLC patients. The training dataset consisted of ST data (Xenium, 10X genomics) and H&E images from 90 NSCLC samples collected from a single institution. After aligning ST data with H&E images, cell-type masks were generated for lymphocyte labeling. For performance evaluation, 456 patches (256 × 256 pixels, 0.45 µm/pixel) were randomly selected from 30 NSCLC H&E slide images. Two pathologists independently annotated lymphocytes in these patches. Of the 456 patches, 355 with consistent annotations between the two pathologists were selected for model evaluation. The consensus annotations served as the reference standard to assess the model's performance in lymphocyte identification. Results: This model achieved an AUROC of 0.8411, indicating high diagnostic accuracy. The optional threshold was determined to be 0.3870, with specificity and sensitivity of 72.76% and 78.92%. Compared to two pathologists’ consensus annotations, the model showed a sensitivity of 81.66%, specificity of 69.70%, and accuracy of 77.29% for lymphocyte identification. Conclusion: This AI model trained from subcellular ST provides a robust performance for analyzing the distribution of lymphocytes in NSCLC H&E images. This model offers an innovative approach to analyzing the composition of cells from H&E images, demonstrating its potential contribution to TME research and the development of precision medicine. Citation Format: Seo Hye Park, Haenara Shin, Hosub Park, Jaemoon Koh, Kwon Joong Na, Hongyoon Choi. Development and validation of an AI-based model for lymphocyte identification in NSCLC H&E image using spatial transcriptomics [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 2421.

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.

Local discriminative representation is needed in many medical image analysis tasks such as identifying sub-types of lesion or segmenting detailed components of anatomical structures. However, the commonly applied supervised representation learning methods require a large amount of annotated data, and unsupervised discriminative representation learning distinguishes different images by learning a global feature, both of which are not suitable for localized medical image analysis tasks. In order to avoid the limitations of these two methods, we introduce local discrimination into unsupervised representation learning in this work. The model contains two branches: one is an embedding branch which learns an embedding function to disperse dissimilar pixels over a low-dimensional hypersphere; and the other is a clustering branch which learns a clustering function to classify similar pixels into the same cluster. These two branches are trained simultaneously in a mutually beneficial pattern, and the learnt local discriminative representations are able to well measure the similarity of local image regions. These representations can be transferred to enhance various downstream tasks. Meanwhile, they can also be applied to cluster anatomical structures from unlabeled medical images under the guidance of topological priors from simulation or other structures with similar topological characteristics. The effectiveness and usefulness of the proposed method are demonstrated by enhancing various downstream tasks and clustering anatomical structures in retinal images and chest X-ray images.KeywordsUnsupervised representation learningLocal discriminationTopological priors

Allergic rhinitis (AR) is a common chronic nasal mucosal inflammatory disorder driven by type 2 immunity, but the spatial stromal-immune cell interactions underlying its pathogenesis remain unclear. We used 10x Genomics Xenium In Situ spatial transcriptomics to map the nasal mucosa of 10 AR patients and 10 non-allergic controls, combined with unsupervised cell clustering, differential gene expression (DE) analysis of COL1A1+PDGFRA+ fibroblasts, qRT-PCR validation, ligand-receptor modeling (CellPhoneDB/NicheNet), and multimodal integration of spatial, transcriptional, and clinical data. Nine major cell types with tissue-specific localization were identified. The AR samples showed expanded fibroblast-rich regions (34.2 ± 3.1% vs. 15.6 ± 2.4% in controls; p<0.001) and increased adjacency between CD4+ T cells and fibroblasts (62.3 ± 4.5% vs. 28.7 ± 3.8% in controls; p<0.001). The fibroblasts in AR had 187 upregulated genes (e.g., TSLP, IL33) that were spatially enriched near CD4+ T cells and validated by qRT-PCR. CD4+ T cells within 20 μm of fibroblasts in AR showed higher Th2 cytokine expression (IL4, IL5, IL13) and Th2/GATA3 signature scores (p<0.001). Three key ligand-receptor axes (TSLP-IL7R, OX40L-OX40, and ICOSL-ICOS) drove the fibroblast-Th2 crosstalk. A "fibroblast-T cell crosstalk score" was ×4.8 higher in AR (p<0.001) and correlated with clinical severity (serum IgE: r=0.71; SPT wheal diameter: r=0.65; p<0.001). AR is defined by expanded fibroblast niches, fibroblast-derived type 2 mediators, and ligand-receptor-dependent fibroblast-Th2 crosstalk-a central pathogenic driver and potential therapeutic target.

Spatial transcriptomics is an emerging technology that allows for analysis of cellular and molecular heterogeneity at spatial resolution. The accurate identification of pathological regions in spatial transcriptomics data is essential for understanding tissue heterogeneity and disease progression. We introduce UPSST, a comprehensive framework that integrates tissue morphology, imputes gene expression, and clusters spatial regions using a graph attention neural network (GAT). UPSST was evaluated across multiple spatial transcriptomics datasets, achieving high performance, such as achieving an Adjusted Rand Index (ARI) of 0.737 and a Fowlkes-Mallows Index (FMI) of 0.818 on slice 151671 of the LIBD human dorsolateral prefrontal cortex (DLPFC) dataset. These results highlight the robustness and precision of our approach in identifying pathology domains. Additionally, UPSST facilitates downstream analyses such as differential and enrichment analysis, which are crucial for deriving biological insights. In conclusion, UPSST offers a powerful and reliable tool for spatial transcriptomics analysis, advancing the identification of pathological regions with high accuracy.

stSCI: A multi-task learning framework for integrative analysis of single-cell and spatial transcriptomics data

ABSTRACTDespite enabling high‐resolution mapping of gene expression within tissues, spatial transcriptomics (ST) still faces challenges in accurately segmenting spatial domains due to complex tissue architecture and limitations of current methods. Most approaches rely on local spatial priors, lack gene‐level interpretability, and fall short in capturing structure‐discriminative genes or long‐range functional relationships, limiting their ability to resolve biologically meaningful architectures. We present Spatially Aware Gene selection and dual‐view Embedding fusion (SAGE), a unified and reproducible framework for domain identification in spatial transcriptomics that combines topic‐driven gene selection with dual‐view embedding fusion to address these gaps. SAGE integrates non‐negative matrix factorization (NMF)‐based topic modeling with classifier‐based importance scoring to identify highly spatially informative genes, and fuses a local expression graph with a topic‐driven non‐local graph via consensus refinement and contrastive graph representation learning to jointly learn spatial and functional embeddings. Evaluated on 34 real‐world datasets, SAGE not only outperforms existing methods in clustering accuracy but also reveals functionally coherent regions and interpretable gene expression patterns. In case studies, SAGE reveals spatial heterogeneity associated with a pre‐malignant activation state in human breast cancer. Moreover, in zebrafish melanoma, it refines the tumor–muscle interface into transcriptionally distinct subdomains and uncovers shared vascular signatures between anatomically separate tissues. Together, these results demonstrate that SAGE can be used not only for accurate spatial domain delineation across diverse ST platforms, but also for dissecting microenvironmental niches and long‐range tissue interactions underlying disease progression.

Spatial transcriptomics (ST) technologies have revolutionized our ability to map gene expression patterns within native tissue context, providing unprecedented insights into tissue architecture and cellular heterogeneity. However, accurately deconvolving cell-type compositions from ST spots remains challenging due to the sparse and averaged nature of ST data, which is essential for accurately depicting tissue architecture. While numerous computational methods have been developed for cell-type deconvolution and spatial distribution reconstruction, most fail to capture tissue complexity at the single-cell level, thereby limiting their applicability in practical scenarios. To this end, we propose a novel cycle-consistent generative adversarial network named STCGAN for cellular deconvolution in spatial transcriptomic. STCGAN first employs a cycle-consistent generative adversarial network (CGAN) to pre-train on ST data, ensuring that both the mapping from ST data to latent space and its reverse mapping are consistent, capturing complex spatial gene expression patterns and learning robust latent representations. Based on the learned representation, STCGAN then optimizes a trainable cell-to-spot mapping matrix to integrate scRNA-seq data with ST data, accurately estimating cellular composition within each capture spot and effectively reconstructing the spatial distribution of cells across the tissue. To further enhance deconvolution accuracy, we incorporate spatial-aware regularization that ensures accurate cellular distribution reconstruction within the spatial context. Benchmarking against seven state-of-the-art methods on five simulated and real datasets from various tissues, STCGAN consistently delivers superior cell-type deconvolution performance. The code of STCGAN can be downloaded from https://github.com/cs-wangbo/STCGAN and all the mentioned datasets are available on Zenodo at https://zenodo.org/doi/10.5281/zenodo.10799113.

Spatial transcriptomics enables spatial gene expression profiling, motivating computational models that capture spatially conditioned regulatory relationships. We introduce SAGE-FM, a lightweight spatial transcriptomics foundation model based on graph convolutional networks (GCN) trained with a masked-central-spot prediction objective. Trained on 416 human Visium samples spanning 15 organs, SAGE-FM learns spatially coherent embeddings that recover masked genes robustly, with 91% of masked genes showing significant correlations (p < 0.05). The SAGE-FM generated embeddings outperform MOFA and spatial transcriptomics in unsupervised clustering and preservation of biological heterogeneity. SAGE-FM generalizes to downstream tasks, enabling 81% accuracy in pathologist-defined spot annotation in oropharyngeal squamous cell carcinoma and improving glioblastoma subtype prediction relative to MOFA. In silico perturbation experiments further show that the model captures directional ligand-receptor and upstream-downstream regulatory effects consistent with ground truth. These results demonstrate that simple, parameter-efficient GCNs can serve as biologically interpretable and spatially aware foundation models for large-scale spatial transcriptomics.

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.

Recent advances of single-cell transcriptomics technologies and allied computational methodologies have revolutionized molecular cell biology. Meanwhile, pioneering explorations in spatial transcriptomics have opened up avenues to address fundamental biological questions in health and diseases. Here, we review the technical attributes of single-cell RNA sequencing and spatial transcriptomics, and the core concepts of computational data analysis. We further highlight the challenges in the application of data integration methodologies and the interpretation of the biological context of the findings.

Understanding local immune responses and the structural integrity of the genital mucosal barrier is essential for advancing knowledge of both susceptibility to sexual HIV transmission and the pathogenesis of chronic HIV infection. This review highlights innovative methodologies for investigating host responses to HIV infection using ex vivo human genital tissues. Recent advances in spatial transcriptomics and multiplex imaging enable high-resolution analysis of tissue architecture, immune cell distribution, and gene expression within mucosal environments. These technologies reveal spatially heterogeneous immune responses as well as epithelial and submucosal structural alterations associated with increased HIV susceptibility and chronic infection. Computational workflows tailored to genital tissue morphology enhance reproducibility and support the integration of transcriptomic and imaging data. Despite current limitations, such as resolution constraints and high costs, ongoing improvements in spatial transcriptomics and bioimaging platforms promise deeper insights into mucosal barrier function and HIV pathogenesis. Characterizing tissue-specific immune dynamics through these approaches may guide the development of targeted interventions aimed at reinforcing mucosal defenses and reducing vulnerability to HIV infection.