Define and visualize pathological architectures of human tissues from spatially resolved transcriptomics using deep learning.
Define and visualize pathological architectures of human tissues from spatially resolved transcriptomics using deep learning.
复制标题
DOI：
10.1016/j.csbj.2022.08.029
复制
发表时间：
2022
影响因子：
6
通讯作者：
Ma, Qin
中科院分区：
文献类型：
--
作者：
Chang, Yuzhou;He, Fei;Wang, Juexin;Chen, Shuo;Li, Jingyi;Liu, Jixin;Yu, Yang;Su, Li;Ma, Anjun;Allen, Carter;Lin, Yu;Sun, Shaoli;Liu, Bingqiang;Otero, Jose Javier;Chung, Dongjun;Fu, Hongjun;Li, Zihai;Xu, Dong;Ma, Qin
空间分辨转录组学提供了一种新的方法来定义空间背景和理解复杂的人类疾病的发病机制。虽然一些计算框架可以通过各种聚类方法来表征空间背景，但由于空间信息关联能力有限，详细的空间结构和功能区划往往无法显示和定位。我们提出了RESEPT，这是一个深度学习框架，用于从空间分辨转录组学中表征和可视化组织结构。给定基因表达或RNA速度等输入，RESEPT通过空间转录组学的空间保留图神经网络学习三维嵌入。然后通过映射到RGB图像中的颜色通道来可视化嵌入，并使用监督卷积神经网络模型进行分割。基于人类和小鼠皮层的10 x Genomics Visium空间转录组学数据集的基准，RESEPT准确地推断和可视化组织结构。值得注意的是，对于内部AD样本，RESEPT可以基于预定义的区域或细胞类型富集基因定位皮质层和细胞类型，并进一步为阿尔茨海默病中β淀粉样蛋白斑块的鉴定提供关键见解。有趣的是，在胶质母细胞瘤样本分析中，RESEPT区分了肿瘤富集的、非肿瘤的和具有浸润性肿瘤细胞的神经胶质瘤区域，以支持临床和预后癌症应用。
Spatially resolved transcriptomics provides a new way to define spatial contexts and understand the pathogenesis of complex human diseases. Although some computational frameworks can characterize spatial context via various clustering methods, the detailed spatial architectures and functional zonation often cannot be revealed and localized due to the limited capacities of associating spatial information. We present RESEPT, a deep-learning framework for characterizing and visualizing tissue architecture from spatially resolved transcriptomics. Given inputs such as gene expression or RNA velocity, RESEPT learns a three-dimensional embedding with a spatial retained graph neural network from spatial transcriptomics. The embedding is then visualized by mapping into color channels in an RGB image and segmented with a supervised convolutional neural network model. Based on a benchmark of 10x Genomics Visium spatial transcriptomics datasets on the human and mouse cortex, RESEPT infers and visualizes the tissue architecture accurately. It is noteworthy that, for the in-house AD samples, RESEPT can localize cortex layers and cell types based on pre-defined region- or cell-type-enriched genes and furthermore provide critical insights into the identification of amyloid-beta plaques in Alzheimer's disease. Interestingly, in a glioblastoma sample analysis, RESEPT distinguishes tumor-enriched, non-tumor, and regions of neuropil with infiltrating tumor cells in support of clinical and prognostic cancer applications.
登录
查看更多内容
影响因子：
6
作者：
Hu J;Schroeder A;Coleman K;Chen C;Auerbach BJ;Li M
通讯作者：
Li M
影响因子：
16.6
作者：
Grauel AL;Nguyen B;Ruddy D;Laszewski T;Schwartz S;Chang J;Chen J;Piquet M;Pelletier M;Yan Z;Kirkpatrick ND;Wu J;deWeck A;Riester M;Hims M;Geyer FC;Wagner J;MacIsaac K;Deeds J;Diwanji R;Jayaraman P;Yu Y;Simmons Q;Weng S;Raza A;Minie B;Dostalek M;Chikkegowda P;Ruda V;Iartchouk O;Chen N;Thierry R;Zhou J;Pruteanu-Malinici I;Fabre C;Engelman JA;Dranoff G;Cremasco V
通讯作者：
Cremasco V
影响因子：
8.8
作者：
Darmanis, Spyros;Sloan, Steven A.;Quake, Stephen R.
通讯作者：
Quake, Stephen R.
影响因子：
64.8
作者：
Eng, Chee-Huat Linus;Lawson, Michael;Cai, Long
通讯作者：
Cai, Long
影响因子：
4.4
作者：
Bergenstrahle, Joseph;Larsson, Ludvig;Lundeberg, Joakim
通讯作者：
Lundeberg, Joakim