Elucidating human beta cell transcriptional regulome with low-input genomic technologies

遗传学研究揭示了许多与糖尿病相关的非编码序列变异 或肥胖风险。生成关键代谢细胞或组织类型的参考 3D 表观基因组数据，例如 因此，人类胰腺β细胞对于了解疾病病因非常重要。 研究转录调控的传统基因组方法包括 RNA-seq 和 ChIP-seq，用于 分别绘制转录组和表观基因组的图谱。这些技术通常需要~1 每次检测百万个细胞。 Hi-C 是目前最流行的 3D 基因组组织图谱方法 这是一种公正的方式，但需要数千万个细胞才能实现千碱基分辨率的 3D 基因组 分析。另一方面，β细胞研究界依赖于死者的胰岛样本 捐助者，这是宝贵且非常昂贵的。因此，低投入技术对于β细胞至关重要 基因组研究。该项目的总体目标是将几种最先进的单细胞和低 输入基因组技术来系统地表征队列中的 β 细胞 3D 增强子调节组 40 个新鲜人类胰岛样本；部分关键技术为实验室原创发明 我们团队的。在目标 1 中，我们将使用大规模并行单细胞 RNA-seq 方法（Drop-seq）来生成 单胰岛细胞转录组的综合数据集，并同时使用低输入 ChIPmentation 方法可绘制 40 名人类胰岛供体队列的增强子和启动子图谱。这个 将导致糖尿病和肥胖症特征基因和可变增强子位点（VEL）的鉴定 与疾病状态相关。在目标 2 中，我们将以千碱基分辨率绘制人类 β 细胞 3D 基因组图谱 使用高效的简易 Hi-C (eHi-C) 方法。地图将揭示之间的所有相互作用 个体增强子和启动子。在目标3中，我们将首次对 人类 INS 基因座的 3D 调节组，并使用 INS 附近的所有增强子的活性进行量化 单倍型解析的人类细胞系。我们还将测试 MAU2-NIPBL 粘连蛋白负载的假设 复合物可能通过介导增强子-启动子 DNA 循环来调节胰岛素基因表达 这个轨迹。最后，我们将使用一种新颖的高通量Mosaic-seq方法来验证体内活性 单细胞水平上数十种β细胞增强剂。这种全面的 3D 调节组数据将提供 了解非编码基因组在 T2D 或肥胖症中的功能的关键资源。

Genetic studies have revealed numerous non-coding sequence variants associated with diabetes or obesity risk. Generating reference 3D epigenome data in in key metabolic cell or tissue types, such as human pancreatic β cell, is therefore very important for the understanding of disease etiology. Conventional genomic methods to study transcriptional regulation include RNA-seq and ChIP-seq, for the mapping of transcriptome and epigenome, respectively. These technologies typically require ~1 million cells per assay. Hi-C is currently the most popular method to map the 3D genome organization in an unbiased fashion, but tens of millions cells are required to achieve kilobase resolution 3D genome analysis. On the other hand, the β cell research community is relying on islet samples from deceased donors, which are precious and very expensive. Low-input technologies are therefore essential for β cell genomic studies. The overall goal of this project is to combine several state-of-art single-cell and low- input genomic technologies to systematically characterize the β cell 3D enhancer regulome from a cohort of 40 fresh human islet samples; some key technologies are the original inventions from the laboratories of our team. In aim 1, we will use a massively parallel single cell RNA-seq method (Drop-seq) to generate a comprehensive dataset of single islet cell transcriptome, and simultaneously use a low-input ChIPmentation method to map enhancers and promoters from a cohort of 40 human islet donors. This will lead to the identification of diabetes and obesity signature genes and variable enhancer loci (VELs) correlated with disease status. In aim 2, we will map the human β cell 3D genome at kilobase resolution using a highly efficient easy Hi-C (eHi-C) method. The map will reveal all the interactions between individual enhancers and promoters. In aim 3, we will for the first time perform an in-depth study of the 3D regulome at the human INS locus, and quantify the activity of all enhancers near INS using a haplotype-resolved human cell line. We will also test a hypothesis that MAU2-NIPBL cohesin loading complex may regulate insulin gene expressions through mediating enhancer-promoter DNA looping at this locus. Finally, we will use a novel high-throughput Mosaic-seq method to validation the in vivo activity of dozens of β cell enhancers at single cell level. This comprehensive 3D regulome data will provide a key resource for the understanding of the functions of non-coding genome in T2D or obesity.