Application of Spatial Omics in the Cardiovascular System

Research, Journal Year: 2025, Volume and Issue: 8

Published: Jan. 1, 2025

Cardiovascular diseases constitute a marked threat to global health, and the emergence of spatial omics technologies has revolutionized cardiovascular research. This review explores application omics, including transcriptomics, proteomics, metabolomics, genomics, epigenomics, providing more insight into molecular cellular foundations disease highlighting critical contributions science, discusses future prospects, technological advancements, integration multi-omics, clinical applications. These developments should contribute understanding guide progress precision medicine, targeted therapies, personalized treatments.

Language: Английский

---

Scaling up spatial transcriptomics for large-sized tissues: uncovering cellular-level tissue architecture beyond conventional platforms with iSCALE

bioRxiv (Cold Spring Harbor Laboratory), Journal Year: 2025, Volume and Issue: unknown

Published: March 1, 2025

Recent advances in spatial transcriptomics (ST) technologies have transformed our ability to profile gene expression while retaining the crucial context within tissues. However, existing ST platforms suffer from high costs, long turnaround times, low resolution, limited coverage, and small tissue capture areas, which hinder their broad applications. Here we present iSCALE, a method that predicts super-resolution automatically annotates cellular-level architecture for large-sized tissues exceed areas of standard platforms. The accuracy iSCALE were validated by comprehensive evaluations, involving benchmarking experiments, immunohistochemistry staining, manual annotation pathologists. When applied multiple sclerosis human brain samples, uncovered lesion associated cellular characteristics undetectable conventional experiments. Our results demonstrate iSCALE's utility analyzing with automatic unbiased annotation, inferring cell type composition, pinpointing regions interest features not discernible through visual assessment.

Language: Английский

---

SIMVI disentangles intrinsic and spatial-induced cellular states in spatial omics data

Nature Communications, Journal Year: 2025, Volume and Issue: 16(1)

Published: March 27, 2025

Spatial omics technologies enable analysis of gene expression and interaction dynamics in relation to tissue structure function. However, existing computational methods may not properly distinguish cellular intrinsic variability intercellular interactions, thus fail reliably capture spatial regulations. Here, we present Interaction Modeling using Variational Inference (SIMVI), an annotation-free deep learning framework that disentangles cell spatial-induced latent variables data with rigorous theoretical support. By this disentanglement, SIMVI enables estimation effects at a single-cell resolution, empowers various downstream analyses. We demonstrate the superior performance across datasets from diverse platforms tissues. illuminates cyclical germinal center B cells human tonsil. Applying multiome melanoma reveals potential tumor epigenetic reprogramming states. On our newly-collected cohort-level CosMx data, uncovers space-and-outcome-dependent macrophage states communication machinery microenvironments. Dissecting properties interactions is crucial for understanding biological processes. authors develop theoretically grounded SIMVI, two factors data.

Language: Английский

---

Virtual RNA Inference from Spatial Transcriptomics Reveals Histology-Associated Pathways that Stratify Metastasis Risk in Colorectal Cancer

medRxiv (Cold Spring Harbor Laboratory), Journal Year: 2025, Volume and Issue: unknown

Published: April 23, 2025

Colorectal cancer (CRC) remains a major health concern, with over 150,000 new diagnoses and more than 50,000 deaths annually in the United States, underscoring an urgent need for improved screening, prognostication, disease management, therapeutic approaches. The tumor microenvironment (TME)-comprising cancerous immune cells interacting within tumor's spatial architecture-plays critical role progression treatment outcomes, reinforcing its importance as prognostic marker metastasis recurrence risk. However, traditional methods TME characterization, such bulk transcriptomics multiplex protein assays, lack sufficient resolution. Although (ST) allows high-resolution mapping of whole transcriptomes at near-cellular resolution, current ST technologies (e.g., Visium, Xenium) are limited by high costs, low throughput, issues reproducibility, preventing their widespread application large-scale molecular epidemiology studies. In this study, we refined implemented Virtual RNA Inference (VRI) to derive ST-level information directly from hematoxylin eosin (H&E)-stained tissue images. Our VRI models were trained on largest matched CRC dataset date, comprising 45 patients 300,000 Visium spots primary tumors. Using state-of-the-art architectures (UNI, ResNet-50, ViT, VMamba), achieved median Spearman's correlation coefficient 0.546 between predicted measured spot-level expression. As validation, VRI-derived gene signatures linked specific regions (tumor, interface, submucosa, stroma, serosa, muscularis, inflammation) showed strong concordance generated via direct ST, performed accurately estimating cell-type proportions spatially H&E slides. expanded cohort controlling invasiveness clinical factors, further identified significantly associated key including status. certain tumor-related pathways not fully captured histology alone, our findings highlight ability infer wide range "histology-associated" biological resolution without requiring profiling. Future efforts will extend framework expand phenotyping standard images, potential accelerate translational research scale.

Language: Английский

---

Deep scSTAR: leveraging deep learning for the extraction and enhancement of phenotype-associated features from single-cell RNA sequencing and spatial transcriptomics data

Briefings in Bioinformatics, Journal Year: 2025, Volume and Issue: 26(3)

Published: May 1, 2025

Abstract Single-cell sequencing has advanced our understanding of cellular heterogeneity and disease pathology, offering insights into behavior immune mechanisms. However, extracting meaningful phenotype-related features is challenging due to noise, batch effects, irrelevant biological signals. To address this, we introduce Deep scSTAR (DscSTAR), a deep learning-based tool designed enhance phenotype-associated features. DscSTAR identified HSP+ FKBP4+ T cells in CD8+ cells, which linked dysfunction resistance checkpoint blockade non-small cell lung cancer. It also enhanced spatial transcriptomics analysis renal carcinoma, revealing interactions between cancer tumor-associated macrophages that may promote suppression affect outcomes. In hepatocellular it highlighted the role S100A12+ neutrophils cancer-associated fibroblasts forming tumor barriers potentially contributing immunotherapy resistance. These findings demonstrate DscSTAR’s capacity model extract phenotype-specific information, advancing mechanisms therapy

Language: Английский

---