Integrative modelling of single-cell data to elucidate the genetic architecture of complex disease

项目摘要/摘要 事实证明，利用基因组广泛关联研究(GWAS)来了解疾病是具有挑战性的，因为 潜在的生物学机制往往很难被块状组织捕捉到。单细胞技术的最新进展 测序导致了跨多个医疗设备、上下文和个人的数据洪流，这提供了 以高分辨率了解疾病生物学的前所未有的机会。我们假设罚金- 由单细胞数据捕获的规模细胞背景将有效地解释疾病的遗传性和细微- 绘制疾病机制图。然而，目前将单细胞数据与GWAs集成的方法在很大程度上 依赖现成的批量测序方法，这掩盖了丰富的表型多样性 存在于规范细胞类型之内和之间的单个细胞中。的稀疏性和高度变化性 单细胞数据也对有效识别单细胞数量性状基因座(QTL)提出了挑战。 单元格数据在大小和复杂性上持续增加，强调需要可扩展的集成式 模特儿。在这里，我们提出了一个5年的研究计划，以开发集成单细胞的新方法 通过对现有批量方法没有捕捉到的复杂细胞表型进行建模，将数据与GWAs进行比较。我们的 提案将确定与疾病相关的新细胞状态；利用多种单细胞模式来精细绘制地图 疾病变异及其靶基因；并发现与疾病相关的新的单细胞QTL。我们的 具体目标是：目标1：利用单细胞表观遗传学数据确定疾病的可遗传成分； 目标2：利用单细胞数据精细绘制疾病变异及其机制图；目标3：定义 利用种群规模的scRNA-seq对疾病变异的调控作用。虽然我们提出的方法是 广泛适用于常见疾病，我们将以免疫相关特征和 在已发表的工作中，我们已经使用大量数据集广泛研究了神经精神病学特征，在那里我们 现在已经汇总了大量相关的单细胞数据集。我们的合作有多个 优势：我们专注于跨多个单细胞医疗设备的功能数据集成；我们广泛的统计数据 和计算专业知识；以及我们关于常见疾病的广泛的、以数据为导向的出版记录。

PROJECT SUMMARY/ABSTRACT Leveraging Genome Wide Association Studies (GWAS) to understand disease has proven challenging, as the underlying biological mechanisms are often poorly captured by bulk tissues. Recent advances in single-cell sequencing have led to a torrent of data across multiple modalities, contexts, and individuals, which provide an unprecedented opportunity to understand disease biology at high resolution. We hypothesize that the fine- scale cellular contexts captured by single-cell data will be effective at explaining disease heritability and fine- mapping disease mechanisms. However, current approaches to integrate single-cell data with GWAS largely rely on off-the-shelf approaches developed for bulk sequencing, which obscure the rich phenotypic diversity present in individual cells within and across canonical cell types. The sparse and highly variable nature of single-cell data has additionally posed challenges for robustly identifying single-cell quantitative trait loci (QTL). Single-cell data continues to increase in size and complexity, emphasizing the need for scalable integrative modeling. Here, we propose a 5 year research plan to develop novel approaches for integrating single-cell data with GWAS by modeling complex cellular phenotypes not captured by existing bulk approaches. Our proposal will identify novel disease-relevant cell states; leverage multiple single-cell modalities to fine-map disease variants and their target genes; and discover novel single-cell QTLs associated with disease. Our specific aims are: Aim 1: Leveraging single-cell epigenetic data to identify heritable components of disease; Aim 2: Leveraging single-cell data to fine-map disease variants and their mechanisms; Aim 3: Defining the regulatory effects of disease variants using population-scale scRNA-seq. While our proposed approaches are broadly applicable to common diseases, we will benchmark them on immune-related traits and neuropsychiatric traits which we have studied extensively with bulk datasets in published work and where we have now aggregated a large collection of relevant single-cell datasets. Our collaboration has multiple strengths: our focus on functional data integration across multiple single-cell modalities; our broad statistical and computational expertise; and our extensive, data-driven publication record on common disease.