Groundwork for a Synchrotron MicroCT Imaging Resource for Biology (SMIRB)

项目摘要 每一种主要的人类疾病都与特定范围的细胞和组织的形态变化有关 微米尺度。正常和异常结构被发现，并仍被组织学表征-a 依靠物理组织切片的显微技术。目前，组织学在系统生物学中的应用是 受其主要是描述性和二维性的限制。使组织学定量化、立体化 将对研究和诊断产生潜在的变革，但一直不切实际。因此，我们 现在已经通过定制固定和染色的X射线显微断层扫描(Micro-CT)创建了组织学的3D形式， 毫米尺度的整个生物体和组织样本。我们用固定的、金属染色的、完整的斑马鱼 因为它们包含了目前组织学研究的大小范围内的各种组织。结果是 这是在任何平面上创建虚拟组织学“切片”的第一种实用方法。立体的、完整的 组织表型在遗传和化学筛查以及临床和毒理学中具有潜在的用途。 组织诊断学。在此，我们提出了实现高吞吐量、量化、3D 整只毫米级动物的组织学表型。拟议的工作应用了以下原则 化学、物理和计算机科学，以提高图像分辨率、吞吐量和分析能力，组织成 三个具体目标。具体目标1将在我们在此项目中的发展基础上再接再厉，进一步改进成像 通过升级成像阵列、光学元件和亚像素移位来提高体积和分辨率，并通过更改来提高吞吐量 在样品嵌入、加载几何和力学、螺旋CT扫描、闪烁体材料和数据方面 通过改进ViewTool基础设施和用户界面实现共享。具体目标2将产生参考 图像以定义正常表型变异的范围并获得与一定范围的潜力相关的样本 申请。具体目标3将把机器学习的能力应用于分割、注释和分析。 总之，这项工作将为大规模遗传和化学筛查奠定实际基础，涉及 毫米尺度，基于三维、定量、组织学表型的整个生物体。仪器仪表 分析将在分辨率、视场、视觉效果、图像质量、 分析潜力、吞吐量、样品稳定性和重现性，并可广泛用于试管和 同步辐射X射线源。体素分辨率将在最大1厘米的视场中至少为0.5μm。 每种细胞类型的表示使图像适合于跨成像模式的交叉参考。 将探索潜在的应用，将开始定义“野生型”，并将训练集自动化 已生成分段。潜在的影响将包括大多数NIH研究所和中心的任务。 全动物基因和化学筛查有望影响药物开发、诊断、 以及我们对基因和环境如何定义表型的基本理解。

Project Summary Each major human disease is associated with a specific range of morphological changes to cells and tissues in the micron scale. Normal and abnormal structure was discovered and is still characterized using histology - a microscopic technique that depends on physical tissue slices. Presently, histology’s use in systems biology is limited by its largely descriptive and two-dimensional nature. Making histology quantitative and three-dimensional would be potentially transformational for research and diagnostics, but has been impractical. Accordingly, we have now created a 3D form of histology by customizing X-ray microtomography (micro-CT) of fixed and stained, millimeter-scale, whole organisms and tissue samples. We used fixed and metal-stained, whole zebrafish because they contain a full range of tissues within the size range currently studied histologically. The result is the first practical way to create virtual histology-like “sections” in any plane. Three-dimensional, complete histological phenotyping has potential use in genetic and chemical screens, and in clinical and toxicological tissue diagnostics. Here, we propose the next steps needed to enable high-throughput, quantitative, 3D histological phenotyping of whole, millimeter-scale animals. The proposed work applies the principles of chemistry, physics, and computer science to improve image resolution, throughput, and analytics, organized into three specific aims. Specific Aim 1 will build on our developments in this project and further improve imaging volume and resolution by upgrading imaging array, optics, and sub-pixel shifting, and to throughput by changes in sample embedding, loading geometry and mechanics, helical CT scanning, scintillator material, and to data sharing by improvements to the ViewTool infrastructure and user interface. Specific Aim 2 will yield reference images to define the range of normal phenotypic variation and to obtain samples related to a range of potential applications. Specific Aim 3 will apply the power of machine learning to segmentation, annotation, and analytics. Together, this work will establish a practical foundation for large-scale genetic and chemical screens involving mm-scale, whole organisms based on 3-dimensional, quantitative, histological phenotyping. The instrumentation and analytics will be state-of-the-art in its combination of resolution, field-of-view, pancellularity, image quality, analytical potential, throughput, sample stability, and reproducibility and largely usable with both tube and synchrotron X-ray sources. The voxel resolution will be at least 0.5 μm across fields-of-view of up to 1 cm. Representation of every cell type make the images suitable for cross-referencing across imaging modalities. Potential applications will be explored, “wild-type” will begin to be defined, and training sets for automated segmentation generated. The potential impact will encompass the missions of most NIH Institutes and Centers. The whole-animal genetic and chemical screens enabled are expected to impact drug development, diagnostics, and our basic understanding of how genes and environment define phenotype.