DMS/NIGMS 1: Topological Study on Histological Images and Spatial Transcriptomics

随着高分辨率转录组学分析技术的快速发展，近年来， 对癌症研究中出现的组织微环境（TM）研究的兴趣增加， 神经科学，即，组织中细胞和结构的集合，如神经组织中的神经元和神经胶质 或肿瘤中的免疫、基质和上皮细胞。这些细胞和结构的空间结构 在组织功能中起着关键作用。研究人员已经获得了高分辨率的转录组学和成像 TM研究数据。这两种信息是相辅相成的。高分辨率转录组学 分析，如单细胞RNA-seq（scRNA），提供细胞水平的分子信息，但不 携带细胞的局部背景信息，而组织学图像分析提供详细的背景，但 不提供相应的细胞基因表达谱。由于缺乏一种新的技术，将它们联合收割机结合起来是一项挑战。 转录组学中的细胞和组织学图像中的细胞之间的一一对应。 该项目旨在统一转录组学和生物图像信息学， TM，将先进的拓扑数据分析（TDA）方法应用于新出现的空间 转录组学（ST）数据。ST数据提供局部空间转录组学。TDA提供了基础 用于研究多组学中丰富的上下文信息。该项目将产生一个空间环境感知 TM转录组学的高分辨率作图。其结果将是非常有影响力的。不仅会 促进正常组织水平的功能表征和机制研究，但也将促进各种 疾病类型的诊断、预后及其机理研究。 PI/Co-PI将创建新的拓扑方法，从细胞中提取丰富的上下文信息 在组织学图像中有多种类型。他们还将提出新的学习算法，以整合此类 将拓扑信息转化为局部ST scRNA数据分析，以更好地区分不同的细胞 类型和状态，建立空间背景和细胞信号基因激活之间的联系，并映射 将转录组学信息转化为整个载玻片图像以用于可视化。

With the rapid advance of high-resolution transcriptomic profiling techniques, recent years have witnessed an increased interest in the study of the tissue microenvironment (TM) arising in cancer research and neuroscience, i.e., the collection of cells and structures in a tissue, such as neuron and glia in neural tissue or immune, stroma, and epithelial cells in tumors. The spatial configuration of these cells and structures plays pivotal roles in tissue function. Researchers have obtained high resolution transcriptomic and imaging data for TM study. The two types of information complement each other. High resolution transcriptomic profiling, such as single-cell RNA-seq (scRNA), provides cellular level molecular information, but does not carry local contextual information of cell, while histology image analysis provides detailed context, but does not provide corresponding cellular gene expression profiles. To combine them is challenging due to a lack of one-to-one correspondence between cells in transcriptomics and cells in histology images. This project intends to unify transcriptomics and bioimage informatics for a comprehensive study of TM, applying the advanced topological data analysis (TDA) methodology on the newly emerged spatial transcriptomic (ST) data. ST data provides localized spatial transcriptomics. TDA provides the foundation for studying rich contextual information in multi-omics. This project will produce a spatial-context-aware high-resolution mapping of TM transcriptomics. The outcome will be highly impactful. It will not only promote normal tissue level functionality characterization and mechanism study, but will also boost various types of diseases’ diagnosis, prognosis as well as their mechanistic studies. The PI/Co-PIs will create new topological approaches to extract rich contextual information from cells of multiple types in histology images. They will also propose new learning algorithms to integrate such topological information into localized ST scRNA data analysis for better differentiation of cells of different types and states, to build connection between spatial context and cell signaling gene activation, and to map transcriptomics information into whole slide image for visualization.