Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “spatial transcriptome”

Search indexed PubMed citations on genomics, clinical trials, systematic reviews and public health. Explore titles, authors and supplied subject terms, then open the PubMed record.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 55 records · Page 3Linked to original sources
PubMed2026

Spatial transcriptomics of Ciona adult brains reveals functional zonalization and insights into neural gland function.

The ascidian Ciona is a pivotal chordate model for illuminating the evolutionary origins of the vertebrate brain. Here, spatial transcriptomics of the adult Ciona neural complex, combined with image-based computational super-resolution mapping, resolved distinct tissue domains including the cerebral ganglion, neural gland, ciliated funnel, neural gland duct/dorsal strand, and body wall muscle. Within the cerebral ganglion, high-resolution mapping revealed clear molecular zonalization separating the cortex and medulla, alongside regional specialization within the cortex itself. The neural gland exhibited localized enrichment of genes associated with extracellular matrix and cell-cell interactions. These spatial features suggest that the neural gland functions as a homeostatic and signaling interface, reminiscent of primitive vertebrate meninges or choroid plexus. Overall, this spatially defined gene expression map provides a foundational framework for understanding functional regionalization in the tunicate brain and its evolutionary relationship to vertebrate nervous systems.

Ciona↗
PubMed2026

Spatial Transcriptomics Identifies Characteristic Immunological Niches in Atopic Dermatitis.

BACKGROUND: Atopic dermatitis (AD) is primarily driven by a Type 2 immune response, with T helper (TH2) cells producing IL-4 and IL-13, thereby promoting inflammation, itch, and a compromised skin barrier. Yet, the spatial organization of pathogenic immune cells and their interactions with stromal and epithelial compartments in human AD skin remain incompletely understood. METHODS: We performed 10× Genomics Visium spatial transcriptomics on FFPE skin biopsies from patients with AD (n = 6), psoriasis (n = 2), and healthy controls (n = 5). Data were integrated with AD single-cell RNA sequencing (scRNA-seq) datasets and complemented by imaging mass cytometry (IMC) and multiplex immunofluorescence (IF) to validate the spatial localization of immune cells. Cell-cell communication analysis revealed putative signaling interactions within immune niches. RESULTS: Spatial clustering resolved tissue compartments and demonstrated transcriptional dysregulation in keratinocytes in AD and psoriasis. AD lesions showed a conserved spatial organization of immune aggregates within the superficial dermis. Integration of scRNA-seq signatures revealed spatially organized co-localization of T cells and mature migratory dendritic cells (mmDCs). We developed a ring-based neighborhood analysis to characterize the cellular organization of the immune-stromal niches, revealing T cell-enriched regions surrounded by inflammatory fibroblasts and activated keratinocytes. Intercellular communication analysis further identified putative signaling within mmDC-T cell niches that may promote pathogenic T cell recruitment and activation. Application of tertiary lymphoid structure (TLS) signatures indicated the presence of TLS-like regions. IMC and IF validated the close spatial proximity between activated TH2 cells and mmDCs. CONCLUSION: AD lesions contain spatially organized TLS-like immune niches at the dermal-epidermal junction, characterized by the close association of T cells and mmDCs and coordinated interactions with surrounding stromal and epithelial compartments. These mmDC-T cell niches may represent potential targets for future therapeutic strategies aimed at disrupting persistent local inflammatory pathways and improving long-term disease control.

atopic dermatitis↗
PubMed2026

Single-cell and spatial transcriptomic technologies for lung cancer tumor microenvironment analysis.

Lung cancer remains one of the leading causes of cancer-related mortality worldwide; beyond its rising incidence, its marked molecular heterogeneity and complex tumor microenvironment (TME) hinder treatment response and drive resistance, contributing directly to its high mortality rate. Single-cell RNA sequencing (scRNA-seq) and spatial transcriptomics (ST) provide complementary approaches for dissecting these features. scRNA-seq enables high-resolution analysis of cellular diversity and transcriptional states but requires tissue dissociation and therefore loses spatial context. In contrast, ST preserves tissue architecture and provides insights into how gene-expression programs within the TME are organized, although no currently available spatial platform combines whole-transcriptome coverage with true single-cell resolution over large tissue areas. Together, these technologies have enabled detailed mapping of tumor, immune and stromal populations and of their spatial interactions, revealing functionally distinct cellular niches that contribute to immune evasion, metastasis and response to therapy. In this narrative review we organize the primary literature around a single question, how spatially structured cellular ecosystems, rather than individual cell types, determine therapeutic response and resistance in lung cancer - and we explicitly separate observations that are reproducible across independent cohorts and platforms from those that remain confined to single studies. We further summarize the technical, analytical and logistic barriers that currently prevent spatially resolved signatures from entering routine diagnostic pathology. Understanding dysregulated pathways and spatially constrained intercellular communication within the TME helps identify candidate biomarkers and may support the identification of therapeutic approaches directed at tumor-intrinsic programs as well as at microenvironment-driven resistance mechanisms.

Cell-cell communication↗
PubMed2025

Colorectal Liver Metastasis Pathomics Model: Integrating Single-Cell and Spatial Transcriptome Analysis With Pathomics for Predicting Liver Metastasis in Colorectal Cancer.

The liver is the primary target organ for hematologic metastasis of colorectal cancer (CRC), and CRC liver metastasis (CRLM) often precludes radical resection, making it the leading cause of death in patients with CRC. To improve the identification and prediction of liver metastasis risk, we identified a cell type of liver metastasis--triggering malignant cells (LMTMCs) through integrating single-cell RNA sequencing and spatial transcriptome analysis. Multiomics cell communication analysis indicated that the interaction between fibroblasts and LMTMCs through the COL1A1-CD44/SDC4 and LAMA4-CD44 signaling axes could promote CRLM. By applying the one-class logistic regression algorithm, we developed a CRLM scoring system in the bulk RNA-sequencing data according to the abundance of LMTMCs in each individual. Using the grouping labels derived from the CRLM scoring system in the bulk data and the corresponding whole-slide images without any manual annotations at the region or pixel level, processed via slide-level weakly supervised learning, a deep-learning model based on the ResNet18 architecture, called Colorectal Liver Metastasis Pathomics Model, was developed to predict the risk of liver metastasis in patients with CRC. The Colorectal Liver Metastasis Pathomics Model achieved an area under the curve of 0.84 at the internal test set of The Cancer Genome Atlas-CRC histology images. In the external independent validation sets, namely the Affiliated Hospital of Southwest Medical University and the Affiliated Traditional Chinese Medicine Hospital of Southwest Medical University cohorts, the areas under the curve were 0.89 and 0.72, respectively, indicating effective classification performances. This study provided new insights and tools for the early identification of CRLM and demonstrated the potential of combining multiomics with deep learning-based pathomics in cancer research.

Humans↗
PubMed2026

Multi-omics and spatial transcriptomics reveal that S100A10 drives CD8+ T-cell exhaustion and immune evasion in hepatocellular carcinoma through cPLA2-5-LOX-mediated arachidonic acid metabolism and ferroptosis.

Immune evasion in hepatocellular carcinoma (HCC) represents a major biological barrier limiting the efficacy of immunotherapy, yet its molecular basis remains incompletely understood. Increasing evidence indicates that tumor metabolic reprogramming and ferroptosis-related signaling play critical roles in shaping an immunosuppressive tumor microenvironment (TME); however, the specific regulatory factors involved remain unclear. This study aims to systematically elucidate the functional role of S100 calcium-binding protein A10 (S100A10) in immune evasion in HCC, with a particular focus on the molecular mechanisms by which S100A10 regulates CD8+ T-cell exhaustion through arachidonic acid (AA) metabolism and ferroptosis, as well as its potential therapeutic implications. To this end, data from The Cancer Genome Atlas Liver Hepatocellular Carcinoma (TCGA-LIHC) cohort are integrated to analyze the expression patterns of S100A10, its prognostic value, and its association with the immune microenvironment. S100A10 overexpression and knockout models are established in HCCLM3 and MHCC97L cell lines, and S100A10-mediated metabolic pathway reprogramming is characterized using transcriptomic profiling, untargeted metabolomics, and ferroptosis-related functional assays. In parallel, single-cell RNA sequencing (scRNA-seq) and spatial transcriptomics are employed to delineate the cell-type specificity and spatial distribution of S100A10. Furthermore, human CD8+ T-cell co-culture systems and orthotopic mouse HCC models are used to evaluate the impact of S100A10 on immune function and responsiveness to anti-programmed cell death protein 1 (anti-PD-1) therapy. The results demonstrate that S100A10 is significantly upregulated in HCC and is closely associated with poor prognosis and an immunosuppressive state. Mechanistically, S100A10 activates cytosolic phospholipase A2-arachidonate 5-lipoxygenase (cPLA2-5-LOX)-mediated AA oxidative metabolism, leading to the accumulation of lipid peroxidation products and ferroptosis-associated signals, thereby driving CD8+ T-cell exhaustion and promoting immune evasion. Significantly, inhibition of S100A10 reshapes the tumor immune microenvironment (TIME) and enhances the therapeutic efficacy of anti-PD-1 treatment. Collectively, these findings identify S100A10 as a critical regulator of metabolic-immune coupling in HCC and provide a theoretical basis for combinatorial strategies targeting metabolism and immunotherapy.

Arachidonic acid metabolism↗
PubMed2026

Integrating histology and spatial transcriptomics via multimodal transformers and contrastive representation learning for accurate gene expression prediction.

Predicting spatial gene expression from Histological images is a fundamental task in understanding tissue organization and molecular phenotypes. However, existing methods often rely on single-model representations or lack effective alignment between image and transcriptomic features. To address these limitations, we propose a unified multimodal learning framework that integrates histological imaging and spatial transcriptomics through a shared latent representation space. Specifically, histological H&E images are encoded by a ResNet50-based convolutional stem and a MobileViT Transformer backbone to extract hierarchical visual representations. Both modalities are projected into a shared latent space via linear-GELU-dropout transformation blocks, enabling cross-modal alignment through a contrastive learning objective that maximizes agreement between the corresponding image and the spot embeddings. Experimental results on the 10x Genomics Visium dataset of human liver tissue demonstrate that MViTGene achieves significantly higher prediction accuracy than existing methods across multiple gene subsets, with improvements of 20%, 33%, and 12% in predicting marker genes, highly expressed genes, and highly variable genes, respectively. The significant improvement in relevance indicates that the model can more accurately capture the true correspondence between tissue morphology and gene expression, therefore enabling more reliable biological interpretation. It provides a computational tool for high-throughput spatial gene expression prediction that balances performance and interpretability.

Humans↗
PubMed2026

A High-Resolution Stereo-Seq Spatial Transcriptomic Resource for Adult Holstein Cattle Liver.

The bovine liver is a highly compartmentalized organ that plays essential roles in continuous gluconeogenesis and nitrogen recycling; however, its spatial molecular architecture has remained largely uncharacterized due to the limitations of traditional bulk and single-cell approaches. To address this gap, Spatial Enhanced Resolution Omics-sequencing (Stereo-seq) was utilized to generate a subcellular-resolution (500 nm) transcriptomic map of an adult Holstein cattle liver, and a refined reference-guided workflow was implemented to overcome standard annotation limitations in livestock. Raw sequencing data were processed using the Stereo-seq Analysis Workflow and analyzed with Stereopy, Seurat, SingleR, and reference-guided workflows. Spatial aggregation was evaluated at Bin20, Bin50, Bin100, Bin150, and Bin200. Increasing bin size increased molecular identifier counts and detected-gene complexity while progressively reducing spatial granularity. Bin50, corresponding to 50 × 50 DNA nanoballs and an approximate nominal footprint of 25 × 25 µm, was therefore selected as a practical intermediate aggregation level for the primary analyses. Quality-control assessment, Leiden clustering, UMAP visualization, reference-based cell-type annotation, cluster-marker analysis, and spatial mapping of canonical hepatic genes demonstrated preservation of biologically interpretable liver transcriptional organization. Raw sequencing data processed spatial matrices, annotated objects, and analysis code are publicly available to support reanalysis and computational benchmarking. In summary, we present a Stereo-seq spatial transcriptomic resource generated from liver tissue of an adult Holstein cow. This initial resource provides a valuable foundation for future studies of bovine liver biology, comparative genomics, and the spatial basis of livestock health and production traits.

Animals↗
PubMed2025

Single-cell-scale spatial transcriptome of the developing and adult mouse ovary.

Mammalian ovary development is essential for female fertility, involving the complex spatial patterning of diverse cell types to establish the finite reserve of ovarian follicles. While single-cell transcriptome analyses have provided important insights into the mechanisms driving specification and developmental trajectories of ovarian cells, they disrupt this crucial spatial context. To overcome this limitation, we used 10X Genomics Visium HD spatial transcriptomics to analyze the developing mouse ovary while maintaining its native cellular architecture. We captured all ovarian cell types at eight key fetal and postnatal timepoints, generating a near single cell resolution library of spatial gene expression across ovarian development. This comprehensive dataset allows analysis of dynamic transcriptional signatures associated with unique spatial patterning throughout development, including the establishment of cortex and medulla and assembly of ovarian follicles in each region. This dataset represents a fundamental resource for the investigation of regulatory mechanisms driving spatial patterning of the ovary and opens new avenues to explore the spatial determinants of female fertility and reproductive longevity.

Journal Article↗
PubMed2025

Spatial transcriptomic analysis of mouse parathyroid gland cells expressing an activating variant of Gcm2.

Glial cells missing 2 (GCM2) is an essential transcription factor for the development of parathyroid glands. Germline GCM2 variants that repress or enhance transcriptional activity predispose a subset of patients to hypoparathyroidism or hyperparathyroidism, respectively. A recurrent germline heterozygous activating missense variant of GCM2, p.Y394S has been identified in some patients with primary hyperparathyroidism. A genetically engineered knock-in mouse model of this variant corresponding to p.Y392S in the mouse Gcm2 gene (Gcm2 +/Y392S) did not show obvious parathyroid tumors. However, in GCM2-binding site mediated luciferase reporter assays in HEK293 cells, the mouse and the human variant both exhibited enhanced transcriptional activity. Therefore, we assessed the effect of this variant on gene expression in vivo in parathyroid glands from Gcm2 +/Y392S and WT mice. Using the 10x Genomics Visium platform, spatially resolved transcriptomic analysis was performed on formalin-fixed and paraffin-embedded (FFPE) tracheal tissue sections of Gcm2 +/Y392S and WT mice to capture RNA from parathyroid glands together with other cell types in the tissue sections. Transcriptome sequence data analysis detected 8 different clusters in the tissue sections based on similarity of gene expression profiles. Cluster-1, which contained parathyroid gland cells expressing Pth and Gcm2, was further evaluated for transcripts that were differentially expressed more than 2-fold in Gcm2 +/Y392S compared to WT. Increased transcript level of Lgals3 (galectin-3) was seen in Gcm2 +/Y392S parathyroid gland cells which is among markers of parathyroid carcinoma. Galectin-3 protein was detected in available FFPE human parathyroid samples of patients with germline heterozygous activating GCM2 variants, p.Y394S (n = 4/10) or p.L379Q (n = 2/2). These results indicate a potential for growth and malignancy of parathyroid glands expressing GCM2 variants. The transcriptomic data of mouse parathyroid gland cells generated in this study can serve as a valuable resource for investigating genes and pathways in normal or abnormal parathyroid gland growth and physiology.

GCM2, gene↗
PubMed2025

Distinct spatial transcriptomic patterns of substantia Nigra in Parkinson disease and Parkinsonian subtype of multiple system atrophy.

To investigate transcriptomic signatures of Parkinson's disease (PD) and the Parkinsonian subtype of Multiple System Atrophy (MSA-P) in substantia nigra pars compacta (SNpc), we conducted transcriptome analysis using in-situ hybridization on paraffin-embedded SNpc tissues from post-mortem brains. The study included 2 MSA-P patients, 2 PD patients, and 2 healthy controls (HC), with 12 regions of interest (ROIs) selected from the dorsal to ventral and medial to lateral aspects of the SNpc. A total of 72 ROIs from 6 participants were analyzed, and differentially expressed genes (DEGs) were identified by comparing MSA-P, PD and HC groups. The MSA-P group showed 88 upregulated DEGs and 326 downregulated DEGs (adjusted &#x1d45d;<0.05) compared to HC. The downregulated DEGs were significantly enriched in pathways related to ribosomal translation, immune processes, mitochondrial function, and autophagy. Notably, the dorsomedial quadrant was uniquely linked to antigen presentation, while other quadrants showed downregulation of protein synthesis. The PD group exhibited 165 upregulated DEGs and 350 downregulated DEGs (adjusted &#x1d45d;<0.05) compared to HC, with downregulated DEGs associated with ribosomal translation, mitochondrial function, and the ubiquitin-proteasome system. In both MSA-P and PD, the upregulated DEGs were not associated with any pathways or biological process in gene enrichment analysis. In network propagation analysis, amyloid precursor protein was the most significant network hub among DEGs in both MSA-P and PD. Comparing the transcriptomic signatures of SNpc between MSA-P and PD, we found immune/inflammation, mitochondrial function and neural signaling related genes were significantly downregulated in MSA-P compared to PD. Overall, the transcriptomic signature of the SNpc in MSA-P and PD revealed overlapping but distinct features, including alterations in protein synthesis, immune processes, mitochondrial function, and protein degradation systems. Future studies with larger cohorts and functional validation are needed to further elucidate these findings.

Humans↗
PubMed2026

Integrative transcriptomic, spatial and functional-genomic analysis identifies a UFMylation-related vascular-stromal program and prioritizes WWTR1 in glioblastoma.

Glioblastoma (GBM) contains spatially organized stress-adaptive and vascular niches. Because transcript abundance does not measure UFM1 conjugation, we asked whether a UFMylation-related transcriptional axis identifies a reproducible tissue program and alters candidate prioritization. In 518 unique primary TCGA-GBM tumors profiled on the Affymetrix HT Human Genome U133A array, weighted gene co-expression network analysis of 8,000 variable genes yielded 12 modules. The 278-gene green module ranked first across nine prespecified traits (mean |r|=0.637). Direct overlap comprised 1/3 measurable UFMylation-core, 5/19 ER-stress/UPR, and 2/15 proteostasis genes; after excluding overlapping genes, correlations with the green eigengene remained significant (r&#x2009;=&#x2009;0.373, 0.831, 0.639, and 0.699 for UFMylation-core, ER-stress/UPR, proteostasis, and composite scores, respectively). The green score was associated with overall survival per standard-deviation increase (HR 1.17, 95% CI 1.07-1.28), although clinical adjustment attenuated the estimate. In a 10-sample single-cell dataset, sample-level scores were higher in pericytes and endothelial cells than in malignant cells. Donor-aware IvyGAP analysis supported regional organization, whereas one Visium section showed stronger concordance with ER-stress/UPR and mesenchymal scores than with the UFMylation-core score. CellChat indicated pathway-selective rather than global remodeling of inferred vascular communication. Layer ablation moved WWTR1 from rank 48 using WGCNA alone to rank 4 overall and rank 1 among non-common-essential genes after cross-platform integration. These findings define an ER-stress/mesenchymal-weighted, UFMylation-related vascular-stromal transcriptional association and nominate WWTR1 for experimental testing.

Humans↗
PubMed2026

Single-cell spatial transcriptomic atlas of the mouse adrenal gland reveals sexual dimorphism in steroidogenic enzyme and hormone receptor expression.

The adrenal cortex shows sexual dimorphism in structure and function. We analysed adrenal glands from 7-week-old male and female BALB/c mice using Visium HD with Cellpose 3 segmentation, comprising 236,077 cells across eleven populations, including four cortical zones. Using curated marker-gene-based zonal annotation, we focused on steroidogenic enzymes and hormone receptors, complementing our companion study based on the same primary dataset. The X-zone was nearly absent in males but prominent in females. Females showed higher Hsd3b1 expression across cortical zones and higher Cyp11b1 expression in outer cortical compartments. The strongest sex difference involved Srd5a2, with markedly higher expression in male zona fasciculata (inner: 77.1% vs. 28.9%), independently supported by RNAscope and immunohistochemistry. Mc2r and Mrap showed discordant spatial distributions, with limited co-expression, suggesting potential MC2R-independent MRAP roles. Agtr1a dominated angiotensin II receptor expression in zona glomerulosa without major sex differences, providing a zone-resolved reference for adrenal sexual dimorphism.

Adrenal cortex↗
PubMed2025

A single-cell spatial transcriptomic census of human skin anatomy.

The skin is the largest human organ and a site of significant disease burden, yet its cellular and molecular organization across the body are largely undefined. Here, we construct a spatially-resolved single-cell atlas of 1.2 million cells from normal adult human skin to localize 45 cell types across 15 anatomic sites. We define principles of organ-wide cell composition, including axes of cell diversity and specialization, and distinguish site-enriched cell types. Each body site is comprised of 10 multicellular neighborhoods that define cell-cell communication. Notably, we identify a perivascular neighborhood enriched for immune-stromal crosstalk with features resembling a homeostatic immune niche similar to skin-associated lymphoid tissue. Finally, mapping these neighborhoods onto skin disease reveals pathogenic neighborhood disruptions, including pan-disease immune alterations in the perivascular neighborhood. We present a framework charting the skin's multiscale spatial organization across a molecular to macroanatomic scale. This work advances our understanding of organ-wide skin cellular organization and communication, and its architectural disruption in disease.

Journal Article↗
PubMed2025

Integrative spatial transcriptomic analysis pinpoints the role of the ferroxidase, TaMCO3, in wheat root tip iron mobilization.

Roots play a critical role in the sensing and absorption of essential minerals from the rhizosphere. Iron (Fe) deficiency, for example, triggers a well-known series of physiological and molecular responses within roots that facilitate uptake, which differs between monocots and dicots. In monocots, little is known about the molecular responses that occur within specific root development zones in response to iron deprivation, and how these differences result in overall nutrient uptake. Here, we conducted a transcriptome analysis of wheat root tips under Fe deficiency (-Fe) and performed a comparative transcriptome analysis with the previous datasets generated from the whole root. Gene ontology analysis of differentially expressed genes highlighted the significance of oxidoreductase activity and metal/ion transport in the root tip, which are critical for Fe mobilization. Interestingly, wheat, an allohexaploid species consisting of three different genomes (A, B, and D) displayed varying gene expression levels arising from the three genomes that contributed to similar molecular functions. Detailed analysis of oxidoreductase function at the root tip revealed multiple multicopper oxidase (MCO) proteins, such as Fe-responsive TaMCO3, that likely contribute to the overall ferroxidase activity. Further characterization of TaMCO3 shows that it complements the yeast FET3 mutant and rescues the -Fe sensitivity phenotype of Arabidopsis atmco3 mutants by enhancing vascular Fe loading. Transgenic wheat lines overexpressing TaMCO3 exhibited increased root Fe accumulation and improved tolerance to -Fe by augmenting the expression of Fe-mobilizing genes. Our findings highlight the role of spatially resolved gene expression in -Fe responses, suggesting strategies to reprogram cells for improved nutrient stress tolerance.

Triticum↗
PubMed2026

Decoding tumor immune microenvironment heterogeneity by single-cell and spatial multi-omics: From immunotherapy resistance to translational biomarkers.

Immune checkpoint blockade has transformed cancer therapy, yet primary and acquired resistance remain major clinical challenges. Increasing evidence indicates that immunotherapy resistance cannot be fully explained by tumor-intrinsic alterations or conventional biomarkers such as PD-L1 expression, tumor mutational burden, or microsatellite instability. Instead, therapeutic response is shaped by the tumor immune microenvironment (TIME) as a heterogeneous, spatially organized, and dynamically evolving ecosystem. Single-cell omics has revealed diverse immune and stromal cell states, including progenitor and terminally exhausted T cells, suppressive myeloid programs, B-cell/TLS-associated immune-reactive states, and CAF-mediated exclusion phenotypes. Spatial transcriptomics, spatial proteomics, and imaging-based approaches further demonstrate that these cell states assemble into distinct immune niches, including immune-inflamed, T-cell-excluded, myeloid-suppressive, metabolic/hypoxic, and TLS-associated niches. These spatial ecosystems determine whether antitumor immune cells can access malignant cells, receive antigen-presenting support, or become restrained by stromal, vascular, metabolic, and myeloid barriers. In this review, we summarize how single-cell and spatial multi-omics redefine TIME heterogeneity in immunotherapy resistance, highlight ligand-receptor communication networks linking cell states to spatial immune dysfunction, and discuss emerging translational biomarkers for patient stratification. We further propose that future immunotherapy biomarkers should evolve from static single-marker assays toward longitudinal, spatially resolved, and interpretable multi-omics models that guide precision combination immunotherapy.

Humans↗
PubMed2026

Representation learning for multi-modal spatially resolved transcriptomics data.

MOTIVATION: Spatial transcriptomics enables in-depth molecular characterization of samples on a morphology and RNA level while preserving spatial location. Integrating the resulting multi-modal data is an unsolved problem, and developing new solutions in precision medicine depends on improved methodologies. RESULTS: We introduce AESTETIK, a convolutional deep learning model that jointly integrates spatial, transcriptomics, and morphology information to learn accurate spot representations. AESTETIK yielded substantially improved cluster assignments on widely adopted technology platforms (e.g. 10x Genomics&#x2122;, NanoString&#x2122;) across multiple datasets. We achieved performance enhancement on structured tissues (e.g. brain) with a 21% increase in median ARI over previous state-of-the-art methods. Notably, AESTETIK also demonstrated superior performance on cancer tissues with heterogeneous cell populations, showing a 2-fold increase in breast cancer, 79% in melanoma, and 21% in liver cancer. We expect that these advances will enable a multi-modal understanding of key biological processes. AVAILABILITY AND IMPLEMENTATION: AESTETIK is implemented in Python 3 and is available as open source software at http://www.github.com/ratschlab/aestetik. The Snakemake pipeline for reproducing the results is available at http://www.github.com/ratschlab/st-rep.

Spatial Transcriptomics↗
PubMed2025

Multi&#x2011;omics identification of a novel signature for serous ovarian carcinoma in the context of 3P medicine and based on twelve programmed cell death patterns: a multi-cohort machine learning study.

BACKGROUND: Predictive, preventive, and personalized medicine (PPPM/3PM) is a strategy aimed at improving the prognosis of cancer, and programmed cell death (PCD) is increasingly recognized as a potential target in cancer therapy and prognosis. However, a PCD-based predictive model for serous ovarian carcinoma (SOC) is lacking. In the present study, we aimed to establish a cell death index (CDI)-based model using PCD-related genes. METHODS: We included 1254 genes from 12 PCD patterns in our analysis. Differentially expressed genes (DEGs) from the Cancer Genome Atlas (TCGA) and Genotype-Tissue Expression (GTEx) were screened. Subsequently, 14 PCD-related genes were included in the PCD-gene-based CDI model. Genomics, single-cell transcriptomes, bulk transcriptomes, spatial transcriptomes, and clinical information from TCGA-OV, GSE26193, GSE63885, and GSE140082 were collected and analyzed to verify the prediction model. RESULTS: The CDI was recognized as an independent prognostic risk factor for patients with SOC. Patients with SOC and a high CDI had lower survival rates and poorer prognoses than those with a low CDI. Specific clinical parameters and the CDI were combined to establish a nomogram that accurately assessed patient survival. We used the PCD-genes model to observe differences between high and low CDI groups. The results showed that patients with SOC and a high CDI showed immunosuppression and hardly benefited from immunotherapy; therefore, trametinib_1372 and BMS-754807 may be potential therapeutic agents for these patients. CONCLUSIONS: The CDI-based model, which was established using 14 PCD-related genes, accurately predicted the tumor microenvironment, immunotherapy response, and drug sensitivity of patients with SOC. Thus this model may help improve the diagnostic and therapeutic efficacy of PPPM.

Humans↗