Data Analysis, Integration, and Sharing Core

项目摘要-数据分析核心 在与骨髓组织测绘核心器官特定项目(OSP)的密切合作下，数据 分析核心(DAC)将监督样本元数据整理、低级别数据处理和质量控制、空间 特征提取、交叉分析整合、组织图谱构建和蜂巢数据传递。我们将利用 经过审查、同行评审的数据管道，用于多路传输组织成像数据的可扩展处理和分析 以前遵循公平的数据共享标准，并已构建为优先考虑互操作性 和透明度。DAC将使用这些工具来构建一个多尺度、多模式的人骨图谱 以年龄、性别、种族和解剖部位为背景的骨髓。高度多路复用的亚细胞成像 使用MIBI-TOF将提供对骨髓结构、细胞类型和功能状态的全面普查 这将与从头开始的空间葡聚糖成像和在相邻的 分别使用MALDI和NanoStringDSP进行连续切片。整体细胞分割和单个细胞分割 通道蛋白表达数据将用于高分辨率MIBI-TOF中的细胞聚类和注释 图像。同时，基于像素的方法将被用来识别细胞中的复杂表型特征- 可以很容易地在MIBI-TOF和MALDI数据上部署的不可知的方式。这些数据将是 关于细胞或基于像素的特征进行询问以识别由以下组成的复合组织邻居 任一特征类型的重复组合。最后，跨每个通道的图像联合配准将允许 用MIBI表征的细胞亚群的空间分布与局部多糖组成和 转录程序。

Project Summary - Data Analysis Core In close collaboration with the bone marrow Tissue Mapping Core Organ Specific Project (OSP), the data analysis core (DAC) will oversee sample metadata curation, low level data processing and quality control, spatial feature extraction, cross-assay integration, tissue atlas construction, and HIVE data delivery. We will leverage vetted, peer-reviewed data pipelines for scalable processing and analysis of multiplexed tissue imaging data previously that adhere to FAIR data sharing standards and have been constructed to prioritize interoperability and transparency. The DAC will use these tools to construct a multiscale, multimodal atlas of human bone marrow that is contextualized by age, gender, race, and anatomic site. Highly multiplexed subcellular imaging using MIBI-TOF will provide a comprehensive census of bone marrow structure, cell types, and functional states that will be coregistered with de novo spatial glycan imaging and tissue transcriptomics performed on adjacent serial sections using MALDI and Nanostring DSP, respectively. Whole cell segmentation together with single channel protein expression data will be leveraged for cell clustering and annotation in high resolution MIBI-TOF images. Meanwhile, pixel-based approaches will be employed to identify complex phenotypic features in a cell- agnostic manner that can be readily deployed on both MIBI-TOF and MALDI data. These data will be interrogated with respect to cell or pixel-based features to identify composite tissue neighborhoods consisting of recurring combinations of either feature type. Lastly, image co-registration across each modality will permit the spatial distributions of cell subsets characterized by MIBI to be correlated with local glycan composition and transcriptional programs.