Decoding the molecular basis of cellular identity in the human brain

项目摘要 了解人脑中细胞特性的分子基础是大脑倡议的一个主要目标 对于阐明各种大脑疾病的细胞起源至关重要。这个项目将使用小说《自上而下》。 补充单细胞和单核方法以确定核心转录的分析策略 在不需要提纯或分离细胞的同时，还可以识别成人大脑中的主要细胞类型。我们的 中心假设是，可以估计细胞类型丰度和转录本之间的协变性。 通过完整组织样本的整合基因共表达分析，该信息可用于 构建定量的细胞类型定义，执行基因表达的数学建模，并识别细胞 生物系统之间特定类型的转录差异。在目标1中，我们将整合细胞类型- 来自&gt；60个数据集和&gt；7000个神经典型成人大脑样本的特定基因共表达模块 确定主要细胞类型的一致转录图谱。这些配置文件将显示哪个主小区 类型主要表达与人类大脑疾病有关的基因，识别新的生物标记物， 并有助于建立评估体外衍生的人类细胞类型对疾病的有效性的“基本事实” 模特儿。我们还将利用高度重复的基因共表达关系来估计细胞ABUN- 舞蹈和开发成人大脑样本中基因表达的预测模型。这些型号将会- 通过可以在新的人脑转录本中测试的具体预测来证明重复性。 包括来自病理样本的那些)；它们还将导致新的分析策略- GES超越了差异表达分析，揭示了相关的微妙转录扰动 与病理学有关。在目标2中，我们将在小鼠身上复制目标1的目标，并实施全面的努力 确定人脑和小鼠脑中主要细胞类型的二元(开/关)表达差异。在《目标3》中， 我们将评估主要中枢神经系统细胞类型转录同一性的区域差异程度。 成人的大脑。预期结果包括星形胶质细胞、少突胶质细胞和 细胞、小胶质细胞、神经元、室管膜细胞和内皮细胞；严格的数学模型可以计算 准确预测人类大脑转录本中数千个基因的表达水平；以及新工具 以及用于研究中枢神经系统细胞类型和亚型的试剂。这个项目是创新的，因为它挑战了 为了研究细胞的分子特性，细胞必须被物理分离的现状；它还介绍了一种新的 用于定义细胞身份的概念和度量称为基因表达保真度。我们的研究将会有积极的 通过为识别区分细胞的转录过程提供前所未有的资源来产生影响 人类大脑区域、物种和疾病状态之间的类型，并将直接有助于我们的长期 目标是从分子数据中构建人类中枢神经系统的全面细胞分类学。

Project Summary Understanding the molecular basis of cellular identity in the human brain is a major goal of the BRAIN Initiative and essential for clarifying the cellular origins of diverse brain disorders. This project will use novel, ‘top-down’ analytical strategies that complement single-cell and single-nucleus methods to define the core transcriptional identities of major cell types in the adult human brain while obviating the need to purify or isolate cells. Our central hypothesis is that covariation between the abundance of cell types and transcripts can be estimated through integrative gene coexpression analysis of intact tissue samples, and this information can be used to construct quantitative cell type definitions, perform mathematical modeling of gene expression, and identify cell type-specific transcriptional differences between biological systems. In Aim 1, we will integrate cell type- specific gene coexpression modules from >60 datasets and >7000 neurotypical adult human brain samples to determine consensus transcriptional profiles of major cell types. These profiles will suggest which major cell types primarily express genes that have been implicated in human brain disorders, identify novel biomarkers, and help to establish the 'ground truth' for assessing the validity of human cell types derived in vitro for disease modeling. We will also leverage highly recurrent gene coexpression relationships to estimate cellular abun- dance and develop predictive models of gene expression in adult human brain samples. These models will im- prove reproducibility through concrete predictions that can be tested in new human brain transcriptomes (in- cluding those from pathological samples) as they become available; they will also lead to new analytical strate- gies that go beyond differential expression analysis to reveal subtle transcriptional perturbations associated with pathology. In Aim 2, we will replicate the goals of Aim 1 in mice and implement a comprehensive effort to identify binary (on/off) expression differences in major cell types between human and mouse brains. In Aim 3, we will assess the extent of regional variation in the transcriptional identities of major CNS cell types in the adult human brain. Expected outcomes include consensus transcriptional profiles of astrocytes, oligodendro- cytes, microglia, neurons, ependymal cells, and endothelial cells; rigorous mathematical models that can accu- rately predict expression levels for thousands of genes in de novo human brain transcriptomes; and new tools and reagents for studying CNS cell types and subtypes. This project is innovative because it challenges the status quo that cells must be physically isolated to study their molecular properties; it also introduces a novel concept and metric called gene expression fidelity for defining cellular identity. Our studies will have a positive impact by providing an unprecedented resource for identifying transcriptional processes that distinguish cell types among human brain regions, species, and disease states, and will contribute directly to our long-term goal of constructing a comprehensive cellular taxonomy of the human CNS from molecular data.