EXPLOITING PATTERNS OF GENE ESSENTIALITY IN HUMAN CELLS TO PREDICT GENE FUNCTION, SYNTHETIC LETHALITY, AND CANCER TARGETS

项目摘要/摘要 必需基因是遗传学和功能基因组学的基础。系统的基因敲除研究 酵母菌定义了细胞增殖所必需的第一套完整的基因，以及随后对 基因重要性因环境和遗传背景的不同而不同，揭示了 功能基因组学：当一个基因在一个基因存在的情况下变得必不可少时，就会出现“合成致命性” 另一个基因的突变或功能丧失，以及在相同生物过程中操作的基因往往 在不同的背景下进行检测时，具有相同的功能丧失表型。 CRISPR/Cas9系统适用于人类，使我们的基因组易于处理，在我的 博士后培训，在我目前担任MD Anderson癌症中心助理教授的职位上，我 在CRISPR筛查方面做出了根本性的贡献。我领导了第一个基因敲除研究，以确定 癌细胞中的核心和上下文特定的基本基因(Hart等人，Cell，2015)，并领导了信息学 将FZD5确定为RNF43突变胰腺癌的特定易损性的工作(Steinhart等人，NAT Med，2017)。我设计了这些研究中使用的所有CRISPR试剂，并随后整合了经验数据 以创建一个更小、更高效的库(TKOv3；可在 Addgene)。我的实验室通过开发分类算法来提升CRISPR信息学的最新水平 基本基因和识别药物-基因相互作用，我们已经定义了金标准的基准 每一项主要筛查研究都采用了必要和非必要基因。 CRISPR在人类细胞中的筛查工作开始取得成果，获得了高质量的数据 从数百个细胞系中分离出来。我们寻求将我们的综合专业知识应用于综合分析和高 通过生物学来探索关于基因重要性在细胞谱系中的变异的问题， 基因和环境。与酵母菌一样，具有相似基因敲除适合度的基因群很可能 参与了相同的生物过程，为破译基因功能提供了一条途径。三分之一的人 蛋白质编码基因是结构性的和不变的表达，但其中一半没有被敲除。 表型。其中许多很可能是由对虾缓冲的，这可能是合成致命相互作用的丰富来源。堆芯 在癌症中，当半合子缺失时，每个细胞都需要的必需品对干扰更敏感 细胞，这可能有助于从第一性原理解释拷贝数重排的适应度约束 癌细胞。在全球范围内，共有的遗传脆弱性模式可能会揭示意想不到的肿瘤亚型， 我们的数据驱动、基于网络的综合分析方法的一个关键目标。最后，我们寻求一种预测性的， 基因重要性的过程级模型，可以解释不同血统和基因的差异，而且 进一步可用于开发减少代表性的CRISPR试剂，该试剂能够实现高信息、低成本 成本筛选方法，用于更有针对性的生物应用。

Project summary/abstract Essential genes are fundamental to genetics and functional genomics. Systematic knockout studies in yeast defined the first complete set of genes essential for cellular proliferation, and subsequent surveys of how gene essentiality varied across environmental and genetic backgrounds revealed foundational principles of functional genomics: that “synthetic lethality” arises when one gene becomes essential in the presence of another gene's mutation or loss of function, and that genes operating in the same biological processes tend to have the same loss-of-function phenotypes when assayed across diverse backgrounds. The adaptation of the CRISPR/Cas9 system to humans has rendered our genome tractable, and in my postdoctoral training and in my current position as Assistant Professor at MD Anderson Cancer Center, I have made fundamental contributions advances in CRISPR screening. I led the first gene knockout study to identify both core and context-specific essential genes in cancer cells (Hart et al., Cell, 2015), and led the informatics effort that identified FZD5 as a specific vulnerability in RNF43-mutant pancreatic cancer (Steinhart et al., Nat Med, 2017). I designed all CRISPR reagents used in these studies, and subsequently integrated empirical data across many published screens to create a much smaller, vastly more efficient library (TKOv3; available on Addgene). My lab has advanced the state of the art in CRISPR informatics by developing algorithms to classify essential genes and to identify drug-gene interactions, and we have defined benchmarks of gold-standard essential and nonessential genes that have been adopted by every major screening study. The CRISPR screening effort in human cells is beginning to bear fruit, with high-quality data available from hundreds of cell lines. We seek to apply our combined expertise in integrative analysis and high- throughput biology to explore questions about the variation in gene essentiality across cellular lineage, genotype, and environment. As with yeast, groups of genes with similar knockout fitness profiles are likely involved in the same biological processes, providing an avenue for deciphering gene function. One-third of all protein-coding genes are constitutively and invariantly expressed, yet half of these show no knockout phenotype. Many are likely buffered by paralogs, potentially a rich source of synthetic lethal interactions. Core essentials, required in every cell, are more sensitive to perturbation when hemizygously deleted in cancer cells, which may help explain from first principles the fitness constraints on copy number rearrangement in cancer cells. Globally, patterns of shared genetic vulnerability are likely to reveal unexpected tumor subtypes, a key goal of our data-driven, network-based integrative analytical approach. Finally, we seek a predictive, process-level model of gene essentiality that can explain variations across lineage and genotype, and that further can be used to develop reduced-representation CRISPR reagents that enable high-information, low- cost screening approaches for more focused biological applications.