Scalable, quantitative, single-cell CRISPR screens

项目总结 许多大型测序计划已经产生了人类遗传多样性的目录。一个主要障碍是 利用这些数据更好地了解健康和疾病是人类基因组的巨大规模：例如， 有大约22,000个蛋白质编码基因，数百万个候选调控元件，以及数千万个 常见的变种。这一挑战推动了日益精细的功能基因组学的发展 使许多不同的假设能够被并行测试的工具。例如，CRISPR筛选文库 使基因组中的每个基因都能被系统地敲除、激活或击倒，提供了一种 测试基因与表型之间联系的方法。虽然非常成功，但这些方法是一体的。 意义相当浪费：量化通常相当于计算介导的sgRNA的表示 这种通过测序产生的干扰，意味着每个基因有数千个细胞，在实践中经常用来确保 统计学上的力量。这个简单的限制意味着，屏幕通常只在少数情况下进行 条件，限制应用于珍贵的细胞类型，如患者样本，并使规模扩大 问题很难解决。在这里，我们利用最近伴随着 普及单细胞RNA测序重温CRISPR筛查任务。我们建议 定量报告测序(QRS)，一种高度可扩展和高度量化的连接方法 CRISPR介导的遗传扰动对池格式的工程表型报告的影响。 使用不需要特殊仪器的简单图书馆准备协议，QRS使CRISPR屏幕能够 单细胞分辨和表型的绝对量化。因此，每个单元格都充当独立的 重复测量，减少单元格输入要求，并支持测量 每种扰动的表型，而不是平均效应。我们通过应用程序演示了这一点 该方法可用于对贵重细胞类型中的弱效应进行高度敏感的筛选，以有效地 量化遗传交互作用(即同时干扰多个基因的紧急效应)，并指导 寻找不完全渗透性表型的驱动因素。该平台将扩展CRISPR中介的 功能基因组学到目前难以解决的环境和细胞类型，并使实验能够 “超越基因组规模。”

PROJECT SUMMARY Numerous large sequencing initiatives have produced catalogs of human genetic diversity. A major obstacle in using these data to better understand health and disease is the vast scale of the human genome: for example, there are ~22,000 protein coding genes, millions of candidate regulatory elements, and tens of millions of common variants. This challenge has motivated the development of increasingly refined functional genomics tools that enable many distinct hypotheses to be tested in parallel. For example, CRISPR screening libraries enable every gene in the genome to be systematically knocked out, activated, or knocked down, providing a means of testing how connect genes to phenotype. Though highly successful, these approaches are in one sense rather wasteful: quantification usually amounts to counting the representation of the sgRNAs mediating the perturbations by sequencing, meaning thousands of cells per gene are often used in practice to ensure statistical power. This simple constraint means that screens are often conducted in only a handful of conditions, limits applications to precious cell types such as patient samples, and makes scaling to larger problems difficult. Here we leverage the explosion of recent innovation that has accompanied the popularization of single-cell RNA sequencing to revisit the task of conducting CRISPR screens. We propose Quantitative Reporter Sequencing (QRS), a highly scalable and highly quantitative approach for connecting CRISPR-mediated genetic perturbations to their effects on an engineered phenotypic reporter in pooled format. Using a simple library prep protocol that requires no special instruments, QRS enables CRISPR screens with single-cell resolution and absolute quantification of phenotype. Each cell therefore serves as an independent replicate measurement, reducing cell input requirements and enabling the measurement of distributions of phenotype for each perturbation rather than average effects. Through applications we demonstrate how this approach can be used to conduct highly sensitive screens for weak effects in precious cell types, to efficiently quantify genetic interactions (i.e. emergent effects of perturbing multiple genes simultaneously), and to guide the search for drivers of incompletely penetrant phenotypes. This platform will extend CRISPR-mediated functional genomics to contexts and cell types that are currently intractable and enable experiments that are “beyond genome-scale.”