arXiv:2404.12973v3 Announce Type: replace-cross 
Abstract: The recent advancement of spatial transcriptomics (ST) allows to characterize spatial gene expression within tissue for discovery research. However, current ST platforms suffer from low resolution, hindering in-depth understanding of spatial gene expression. Super-resolution approaches promise to enhance ST maps by integrating histology images with gene expressions of profiled tissue spots. However, current super-resolution methods are limited by restoration uncertainty and mode collapse. Although diffusion models have shown promise in capturing complex interactions between multi-modal conditions, it remains a challenge to integrate histology images and gene expression for super-resolved ST maps. This paper proposes a cross-modal conditional diffusion model for super-resolving ST maps with the guidance of histology images. Specifically, we design a multi-modal disentangling network with cross-modal adaptive modulation to utilize complementary information from histology images and spatial gene expression. Moreover, we propose a dynamic cross-attention modelling strategy to extract hierarchical cell-to-tissue information from histology images. Lastly, we propose a co-expression-based gene-correlation graph network to model the co-expression relationship of multiple genes. Experiments show that our method outperforms other state-of-the-art methods in ST super-resolution on three public datasets.

تُحدث التطورات الأخيرة في علم النسخ الجيني المكاني ثورة في فهمنا للتعبير الجيني في الأنسجة. ومع ذلك، تواجه المنصات الحالية تحديات بسبب انخفاض الدقة. تهدف تقنيات الدقة الفائقة إلى تحسين هذه الخرائط من خلال دمج صور علم الأنسجة مع بيانات التعبير الجيني، مما يمهد الطريق لفهم أعمق لديناميات الجينات المكانية.

Los recientes avances en transcriptómica espacial están revolucionando nuestra comprensión de la expresión génica en los tejidos. Sin embargo, las plataformas existentes enfrentan desafíos debido a la baja resolución. Las técnicas de superresolución buscan mejorar estos mapas al combinar imágenes de histología con datos de expresión génica, allanando el camino para obtener una comprensión más profunda de la dinámica génica espacial.

Les récentes avancées en transcriptomique spatiale révolutionnent notre compréhension de l'expression génique dans les tissus. Cependant, les plateformes existantes rencontrent des défis liés à la faible résolution. Les techniques de super-résolution visent à améliorer ces cartes en combinant des images d'histologie avec des données d'expression génique, ouvrant la voie à des aperçus plus profonds des dynamiques géniques spatiales.

Recent advancements in spatial transcriptomics are revolutionizing how we understand gene expression in tissues. However, existing platforms face challenges with low resolution. Super-resolution techniques aim to improve these maps by combining histology images with gene expression data, paving the way for deeper insights into spatial gene dynamics.

Cross-modal Diffusion Modelling for Super-resolved Spatial Transcriptomics

arXiv:2509.09923v2 Announce Type: replace-cross 
Abstract: Differentiating between the two main subtypes of Inflammatory Bowel Disease (IBD): Crohns disease (CD) and ulcerative colitis (UC) is a persistent clinical challenge due to overlapping presentations. This study introduces a novel computational framework that employs spatial transcriptomics (ST) to create an explainable machine learning model for IBD classification. We analyzed ST data from the colonic mucosa of healthy controls (HC), UC, and CD patients. Using Non-negative Matrix Factorization (NMF), we first identified four recurring cellular niches, representing distinct functional microenvironments within the tissue. From these niches, we systematically engineered 44 features capturing three key aspects of tissue pathology: niche composition, neighborhood enrichment, and niche-gene signals. A multilayer perceptron (MLP) classifier trained on these features achieved an accuracy of $0.774 \pm 0.161$ for the more challenging three-class problem (HC, UC, and CD) and $0.916 \pm 0.118$ in the two-class problem of distinguishing IBD from healthy tissue. Crucially, model explainability analysis revealed that disruptions in the spatial organization of niches were the strongest predictors of general inflammation, while the classification between UC and CD relied on specific niche-gene expression signatures. This work provides a robust, proof-of-concept pipeline that transforms descriptive spatial data into an accurate and explainable predictive tool, offering not only a potential new diagnostic paradigm but also deeper insights into the distinct biological mechanisms that drive IBD subtypes.

تم تقديم إطار حسابي جديد لتمييز بين مرض كرون والتهاب القولون التقرحي، وهما النوعان الرئيسيان من مرض الأمعاء الالتهابي (IBD). يستخدم هذا الإطار النسخ المتماثل المكاني لتطوير نموذج تعلم آلي قابل للتفسير، محققًا دقة تصنيف تبلغ 0.774 لمرضى IBD بناءً على تحليل الفجوات الخلوية.

Se ha introducido un nuevo marco computacional para diferenciar entre la enfermedad de Crohn y la colitis ulcerosa, dos subtipos principales de la enfermedad inflamatoria intestinal (EII). Este marco utiliza la transcriptómica espacial para desarrollar un modelo de aprendizaje automático explicable, logrando una precisión de clasificación de 0.774 para pacientes con EII basado en el análisis de nichos celulares.

Un nouveau cadre computationnel a été introduit pour différencier la maladie de Crohn et la colite ulcéreuse, deux sous-types principaux de la maladie inflammatoire de l'intestin (MII). Ce cadre utilise la transcriptomique spatiale pour développer un modèle d'apprentissage automatique explicable, atteignant une précision de classification de 0,774 pour les patients atteints de MII sur la base de l'analyse des niches cellulaires.

A novel computational framework has been introduced to differentiate between Crohn's disease and ulcerative colitis, two main subtypes of Inflammatory Bowel Disease (IBD). This framework utilizes spatial transcriptomics to develop an explainable machine learning model, achieving a classification accuracy of 0.774 for IBD patients based on cellular niche analysis.

Cross-modal Diffusion Modelling for Super-resolved Spatial Transcriptomics

Was this article worth reading? Share it

LucidQuery AI

OneSky Localization Agent

SVGenius

Deptho.ai

Transcend

Sellm

Ready to build your own newsroom?