Integrated analysis of multi-omic QTLs at single cell resolution

项目摘要 新的统计和计算工具使基因组学技术得以广泛采用， 作为现代人类基因组学的基础。事实上，最近单细胞的出现和普及 基因组学平台--最著名的是单细胞RNA测序（scRNA-seq）--已经导致了单细胞RNA的增殖， 单元数据处理、QC和分析框架。然而，要完全实现单细胞基因组学的承诺， 我们需要将细胞类型水平的调节表型与复杂疾病联系起来。最有前途的 应对这一挑战的方法是通过定位数量性状基因座来鉴定功能相关的遗传变异 （QTL）。事实上，鉴定与许多调节表型相关的调节变异的研究， 包括但不限于基因表达（eQTL）、DNA甲基化（meQTL）、染色质可及性 （caQTL）和蛋白质（pQTL），已经在批量样品中广泛进行。这些研究 推进了我们对复杂疾病的分子基础的理解，但缺乏粒度 大量分析所提供的数据继续阻碍进展。当我们将这些方法转向细胞类型时 通过单细胞基因组学实现的水平分析已经清楚地表明， 不太适合处理单细胞数据的复杂性和特定特征;或者实际上不适合全面 单细胞数据的分辨率和丰富性的优势。在这里，我们建议开发、验证和 利用单细胞组学技术的数据，开发QTL定位方法。我们认为， 重要的是使用从处于疾病状态的原代人体组织获得的相关数据来建立这些方法。 因此，我们将联合收集来自两种组织类型的scRNA-seq、scATAC-seq和单细胞蛋白水平： 以及从肺纤维化（PF）患者或健康对照收集的外周血。此数据收集 将通过我们现有的生物储存库促进，并建立在我们的专业知识基础上，用于分析基因组 数据，绘制调控表型的QTL，并分析从肺组织收集的scRNA-seq， 使用这些数据，我们将建立单变量scQTL定位、多组学scQTL定位和多基因组学scQTL定位的方法。 定位、背景特异性scQTL的鉴定以及scQTL结果与来自 GWAS研究。这些方法将作为开源软件包发布， 外地我们的团队汇集了统计基因组学、计算生物学、功能基因组学、基因组学 基因组学、单细胞基因组学和PF中的疾病特异性专业知识使我们特别适合携带 出这项工作。

Project Summary Novel statistical and computational tools have enabled the broad adoption of genomics technologies and served as the foundation for the modern age of human genomics. Indeed, the recent advent and popularity of single cell genomics platforms – most notably single cell RNA sequencing (scRNA-seq) – has led to a proliferation of single- cell data processing, QC, and analysis frameworks. However, to fully realize the promise of single cell genomics approaches we need to connect cell-type level regulatory phenotypes with complex disease. The most promising approach to this challenge is to identify functionally relevant genetic variation by mapping quantitative trait loci (QTLs). Indeed, studies identifying regulatory variation associated with a number of regulatory phenotypes, including but not limited to gene expression (eQTLs), DNA methylation (meQTLs), chromatin accessibility (caQTLs), and protein (pQTLs), have been carried out extensively in bulk samples. These studies have advanced our understanding of the molecular underpinnings of complex disease, but the lack of granularity provided by bulk analyses continues to hinder progress. As we move these approaches towards the cell-type level analyses enabled by single cell genomics it has become clear that the methods developed for bulk samples are not well suited to handle the complexity and specific characteristics of single cell data; or indeed to take full advantage of the resolution and richness of single cell data. Here we propose developing, validating, and deploying methods for mapping QTLs using data from single cell `omics technologies. We believe it is critically important to build these methods using relevant data obtained from primary human tissue in a disease state. Thus, we will jointly collect scRNA-seq, scATAC-seq and single cell protein levels from two tissue types: lung and peripheral blood collected from patients with pulmonary fibrosis (PF) or healthy controls. This data collection will be facilitated by our existing biorepository and build upon our expertise building tools for analyzing genomic data, mapping QTLs for regulatory phenotypes, and analyzing scRNA-seq collected from lung tissue from patients with PF. Using these data we will build methods for univariate scQTL mapping, multi-omic scQTL mapping, the identification of context specific scQTLs, and integration of scQTL results with the results from GWAS studies. These methods will be released as open-source software packages enabling broad adoption by the field. Our team brings together unique expertise in statistical genomics, computational biology, functional genomics, single cell genomics, and disease specific expertise in PF making us particularly well suited to carry out this work.