Scalable technologies for brain-wide connectomics of transcriptomic cell types: focus on brainstem

项目摘要，用于脑全连接组学的可扩展技术， 转录组细胞类型：聚焦脑干 该建议是开发一个可扩展的管道，以将联合收割机高分辨率形态学和分子生物学相结合。 分类单个神经元以定义脑中的形态分子细胞类型。完整形态学 单个神经元提供了大脑中连接和信息处理的见解，并揭示了如何 神经元活动在大脑区域中传递。将转录组信息分层到形态上不同的 类型提供了访问这些定义的神经元类型以进行功能分析的基础。这样的组合 大脑细胞类型的分类是理解神经元类型在大脑中的作用的基础。 神经回路以及多区域神经回路内的信息处理如何协调复杂的 行为。 基于测序的方法已被用于将脑细胞分类为转录组类型（t- 类型）具有高吞吐量。全脑形态类型完整描述的并行策略 这两种模式的综合分析速度太慢，而且缺乏对这两种模式进行综合分析的方法。为了解决这个问题，我们将 创建全脑成像和神经元重建平台，使用选择性 平面照明显微镜和基于现代机器学习工具的加速重建 U-Nets和强化学习我们将联合收割机与转录后组学表征相结合， 用多重荧光原位杂交重建细胞以确定形态分子类型。 我们将从一个关键的大脑区域，髓质，创建一个包含2,000个这样的双重分类神经元类型的数据集， 鼠标髓质是由不同类型的神经元组成的，这些神经元分布在许多相互关联的核团中 对自主神经功能如呼吸、血管控制、整合来自 感觉和内感受通道以及运动动作的协调，如咀嚼、舔和吞咽。 我们的形态分子延髓神经元类型普查将奠定一个系统的细胞类型的基础 在脑干回路中以及在更大的背景下， 局部脑回路此外，这将作为在整个 老鼠的大脑和其他大脑，包括灵长类动物的大脑。

Project summary, Scalable technologies for brain-wide connectomics of transcriptomic cell types: focus on brainstem This proposal is to develop a scalable pipeline to combine high-resolution morphology and molecular classification of individual neurons to define morpho-molecular cell types in the brain. Complete morphology of individual neurons provides insights of connectivity and information processing in the brain and reveals how neuronal activity is routed across brain areas. Layering transcriptomic information on to morphologically distinct types provides the basis to access these defined neuron types for functional analysis. Such a combined classification of the brain’s cell types is foundational for understanding the role of defined neuron types within neural circuits and how information processing within multi-regional neural circuits orchestrate complex behaviors. Sequencing-based approaches have been used to categorize the brain’s cells into transcriptomic types (t- types) with high throughput. Parallel strategies for a complete description of the morphological types brain-wide are too slow and methods for a combined analysis of these two modalities are lacking. To address this, we will create a brain-wide imaging and neuronal reconstruction platform that provides faster imaging using selective plane illumination microscopy and accelerated reconstructions with modern machine learning tools based on U-Nets and reinforcement learning. We will combine this with post hoc transcriptomic characterization of reconstructed cells with multiplexed fluorescent in situ hybridization to define morpho-molecular types. We will create a data set of 2,000 such dual categorized neuron types from a critical brain area, the medulla, in the mouse. The medulla is comprised of diverse neuronal types organized in numerous inter-related nuclei essential for autonomic functions such as breathing, vasomotor control, integration of ascending inputs from sensory and interoceptive channels and coordination of motor actions such as chewing, licking and swallowing. Our census of morpho-molecular medullary neuron types will lay the foundation for a systematic cell type specific functional interrogation of these neurons within brainstem circuits and in the larger context of multi- regional brain circuits. Furthermore, this will serve as the blueprint for carrying out such studies throughout the mouse brain, and other brains, including that of primates.