美国康奈尔大学王光煜团队的最新研究提出了将组织病理学与空间转录组学连接起来的视觉组学基础模型。相关论文于2025年5月29日发表在《自然—方法学》杂志上。

为了解决这个问题，研究组开发了OmiCLIP，这是一个视觉组学基础模型，将苏木精和伊红图像与来自Visium数据的转录组学主题组织斑块连接起来。该课题组通过连接来自每个片段的顶部表达基因符号，将转录组数据转换为“句子”。研究组策划了一个包含220万对组织图像和转录组数据的数据集，涵盖32个器官，以训练OmiCLIP整合组织学和转录组学。

在OmiCLIP的基础上，他们的Loki平台提供了五个关键功能：组织比对，通过大量RNA测序或标记基因进行注释，细胞类型分解，图像转录组检索以及从苏木精和伊红染色图像中预测空间转录组基因表达。与22个最先进的模型、5个模拟、19个公开和4个室内实验数据集相比，Loki显示出一致的准确性和灵活性。

据了解，人工智能彻底改变了计算生物学。组学技术的最新发展，包括单细胞RNA测序和空间转录组学，提供了详细的基因组数据和组织组织学。然而，目前的计算模型要么集中在组学上，要么集中在图像分析上，缺乏它们的集成。

附：英文原文

Title: A visual–omics foundation model to bridge histopathology with spatial transcriptomics

Author: Chen, Weiqing, Zhang, Pengzhi, Tran, Tu N, Xiao, Yiwei, Li, Shengyu, Shah, Vrutant V., Cheng, Hao, Brannan, Kristopher W., Youker, Keith, Lai, Li, Fang, Longhou, Yang, Yu, Le, Nhat-Tu, Abe, Jun-ichi, Chen, Shu-Hsia, Ma, Qin, Chen, Ken, Song, Qianqian, Cooke, John P., Wang, Guangyu

Issue&Volume: 2025-05-29

Abstract: Artificial intelligence has revolutionized computational biology. Recent developments in omics technologies, including single-cell RNA sequencing and spatial transcriptomics, provide detailed genomic data alongside tissue histology. However, current computational models focus on either omics or image analysis, lacking their integration. To address this, we developed OmiCLIP, a visual–omics foundation model linking hematoxylin and eosin images and transcriptomics using tissue patches from Visium data. We transformed transcriptomic data into ‘sentences’ by concatenating top-expressed gene symbols from each patch. We curated a dataset of 2.2 million paired tissue images and transcriptomic data across 32 organs to train OmiCLIP integrating histology and transcriptomics. Building on OmiCLIP, our Loki platform offers five key functions: tissue alignment, annotation via bulk RNA sequencing or marker genes, cell-type decomposition, image–transcriptomics retrieval and spatial transcriptomics gene expression prediction from hematoxylin and eosin-stained images. Compared with 22 state-of-the-art models on 5 simulations, and 19 public and 4 in-house experimental datasets, Loki demonstrated consistent accuracy and robustness.

DOI: 10.1038/s41592-025-02707-1

Source: https://www.nature.com/articles/s41592-025-02707-1