Evolutionary Human Genomics: Demography, Natural Selection, and Transcriptional Regulation

项目摘要 我的研究计划旨在通过分子进化的镜头来理解现代基因组数据。 借鉴统计学、人工智能和种群遗传学的思想和技术，我们寻求两者 了解形成当今基因组序列的进化力量，并使用进化 洞察特定基因组序列的表型重要性的模式，具有广泛的含义 为了人类健康。我们的工作主要集中在三个方面：(1)基于 基因组序列；(2)人类突变适合性结果的推断；以及(3)对 哺乳动物的转录调控及其进化。 在上一个供资期间(2018-2022年)，我们在这两个领域的每一个领域都取得了重大进展，包括 新方法在重要和及时的科学问题上的方法进步和应用。为 例如，我们最近开发了新的方法来推断祖先重组图 (ARGS)来自多种群序列数据；用于使用ARGS和深度神经网络检测选择性扫描 网络；用于从CRISPR-Cas9筛选中鉴定必需基因；用于相对RNA的估计 基于广泛可用的RNA测序数据类型的半衰期；以及用于检测 基于表观基因组数据的沿系统发育分支的顺式调控元件。这些方法都有 已作为公开可用的软件工具实施。基于这些方法和其他方法，我们发布了一个 各种新的科学发现，包括以前未知的发现和表征 现代和古人类之间的相互渗透事件；南美鸟类的基因组分析 表明它们的辐射主要是由最近的选择性扫描驱动的；一项分析表明 蝙蝠免疫和癌症相关基因快速进化的证据；以及一项分析表明 在哺乳动物进化中，增强子的获得和丢失的速度大约是启动子的两倍。这些发现是 在总共17篇原创论文和预印本中进行了描述。 对于这次续期申请，我们建议继续在这三个关键领域中的每一个领域进行研究。特定的 目标包括开发新的统计抽样方法，以扩展到非常大的ARG；使用域 适应以减少种群基因组学中的训练偏差；测量从 人类种群中罕见的变异模式；多基因适应度效应分布的特征 特征；识别和表征与优势等位基因相关联的有害变异；以及制定统一的 新生RNA测序数据的生物物理建模框架及其在比较基因组学中的应用 延伸率估算。R35资金的续期将使我们能够保持高生产率的捐赠者 这一极其重要的研究领域。

Project Summary My research program aims to make sense of modern genomic data through the lens of molecular evolution. Drawing from ideas and techniques in statistics, artificial intelligence, and population genetics, we seek both to understand the evolutionary forces that have shaped present-day genome sequences, and to use evolutionary patterns to gain insight into the phenotypic importance of particular genomic sequences, with broad implications for human health. Our work focuses in particular on three major areas: (1) evolutionary reconstruction based on genome sequences; (2) inference of the fitness consequences of human mutations; and (3) the study of transcriptional regulation and its evolution in mammals. In the last funding period (2018–2022), we achieved major advances in each of these areas, including both methodological advances and applications of new methods to important and timely scientific questions. For example, we recently developed innovative new methods for the inference of ancestral recombination graphs (ARGs) from multi-population sequence data; for the detection of selective sweeps using ARGs and deep neural networks; for the identification of essential genes from CRISPR-Cas9 screens; for the estimation of relative RNA half-lives based on widely available RNA-sequencing data types; and for the detection of gains and losses of cis-regulatory elements along the branches of a phylogeny based on epigenomic data. These methods have all been implemented as publicly available software tools. Based on these and other methods, we published a variety of novel scientific findings, including the discovery and characterization of previously unknown introgression events between modern and archaic hominins; a genomic analysis of South American birds indicating their radiation was primarily driven by recent selective sweeps; an analysis indicating extensive evidence of rapid evolution in immunity- and cancer-related genes in bats; and an analysis indicating that enhancers are gained and lost at about twice the rate of promoters in mammalian evolution. These findings were described in a total of 17 original papers and preprints. For this renewal application, we propose to continue our research within each of these three key areas. Specific goals include developing new statistical sampling methods that scale to very large ARGs; using domain adaptation to reduce training bias in population genomics; measuring extreme levels of purifying selection from patterns of rare variation in human populations; characterizing the distributions of fitness effects for polygenic traits; identifying and characterizing deleterious variants linked to advantageous alleles; and developing a unified biophysical modeling framework for nascent RNA sequencing data with applications to comparative genomics and elongation-rate estimation. A renewal of R35 funding will enable us to remain highly productive contributors to this critically important research area.