The Architecture of Missing and Archaic Variation in Human Population Genomic Data

人类基因组数据中缺失和古老变异的结构

• 批准号: 10292375
• 负责人: Kimberly Ayers
• 金额: $ 44.1万
• 依托单位: CALIFORNIA STATE UNIVERSITY SAN MARCOS
• 依托单位国家: 美国
• 财政年份: 2021
• 资助国家: 美国
• 起止时间: 2021-07-01 至 2024-06-30
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/10292375
• 关键词: Accounting; Address; Admixture; Africa; African; Algorithms; Alleles; Architecture; Asia; Asians; Benchmarking; Bioinformatics; Chinese People; Clinical; Computer software; Computing Methodologies; Data; Development; Disease; Environment; European; Event; Evolution; Gene Frequency; Genes; Genome; Genomics; Genotype; Health; Human; Human Genome; Individual; Joints; Lead; Link; Linkage Disequilibrium; Markov chain Monte Carlo methodology; Memory; Methods; Modeling; Modern 1601-history; Modernization; Morphologic artifacts; Mosaicism; Muscle; Natural Selections; Oceania; Pattern; Phenotype; Population; Population Sizes; Process; Recording of previous events; Relaxation; Sampling; Series; Shapes; Source; Statistical Methods; Structure; Techniques; Testing; Time; Training; Utah; Variant; Work; base; biomedical data science; career; climate change; diverse data; expectation; functional genomics; genetic variant; genome-wide; genomic data; genomic locus; genomic variation; graduate student; human model; migration; models and simulation; novel; population genetic structure; programs; software development; structural genomics; tool; trait; undergraduate student

Project Summary Modern human genomes are mosaics of variation from numerous archaic non-human hominins, often termed “ghost” populations. However, our understanding of the evolutionary history of “ghost” variation is still developing. Importantly, computational methods to address missing “ghost” variation are still nascent, and not accounting for the presence of “ghosts” often leads to erroneous inference. Here I propose a series of programmatic developments to address inference of evolutionary history from modern human genomes, while accounting for gene flow from archaic “ghosts”. In AIM 1, I propose to develop a parallelized statistical framework for estimating population genetic structure from multi-allelic, multi-locus genomic data that incorporates sequencing and imputation errors of data considered missing due to gene flow from archaic “ghost” populations into a maximum likelihood based statistical framework. This method will be incorporated into a computationally efficient program called p-MULTICLUST, a multi-threaded, parallelized tool which extends the popular “admixture” model incorporated in tools like STRUCTURE and ADMIXTURE to account for missing multi- allelic human genomic data. AIM 2 will involve a two-pronged approach to estimate evolutionary history and population structure in the presence of gene flow from an archaic “ghost” under the Isolation with Migration (IM) model. We will (a) develop extensions to the IMa3/IMa2p suite of tools to incorporate joint estimation of population structure and demographic history from genomic data, and (b) train undergraduate students in developing simulation models for the stdpopsim consortium under two important models of human history – (1) archaic “ghost” gene flow in native Africans, and (2) multiple-epochs of admixture into Asians/Oceanians. In AIM 3, I propose to quantify the selection landscape of “ghost” variation across diverse human genomes due to ancestral gene flow from now extinct “ghost” populations. In this aim, we will focus on (a) improvements to the MigSelect program to quantify linked selection effects due to gene flow from “ghost” populations under the IM model, and (b) a larger, more encompassing study of functional genomic variation across diverse human populations including high-quality genomes from Africa, supplemented with more complete Neanderthal, and other non-human hominin genomes which will help us delineate patterns of human evolutionary history, and understand the functional consequences of archaic gene flow. These discoveries also have direct consequences for understanding modern human ancestry, and disease allele evolution. Importantly, this R15 will train numerous underrepresented Undergraduate and Graduate students in genomics and bioinformatics, towards careers in the biomedical and data sciences.

项目摘要 现代人类基因组是许多古老的非人类古人类基因组变异的镶嵌体， “幽灵”人口。然而，我们对“幽灵”变异进化史的理解仍在发展中。 重要的是，解决缺失的“幽灵”变异的计算方法仍然是新生的， “鬼”的存在往往会导致错误的推论。在这里，我提出了一系列的纲领性建议。 发展，以解决从现代人类基因组推断进化历史，而 解释来自远古“鬼魂”的基因流在AIM 1中，我建议开发一个并行统计 从多等位基因、多基因座基因组数据估计群体遗传结构的框架， 合并了由于来自古代“幽灵”的基因流而被认为缺失的数据的测序和插补错误 最大似然统计框架。该方法将被纳入一个 名为p-MULTICLUST的计算高效程序，这是一个多线程、并行化工具，它扩展了 流行的“混合”模型被合并到像STRUCTURE和ADMIXTURE这样的工具中，以解决缺失的多个 等位基因人类基因组数据。AIM 2将涉及一种双管齐下的方法来估计进化历史， 在隔离与迁移（IM）下，存在来自古老“幽灵”的基因流的种群结构 模型我们将（a）开发IMA 3/IMA 2 p工具套件的扩展，以纳入联合估计， 从基因组数据的人口结构和人口统计学历史，以及（B）在以下方面训练本科生 根据人类历史的两个重要模型，为stdpopsim联合会开发模拟模型-（1） 古老的“幽灵”基因流在非洲土著人，和（2）多个时代的混合到亚洲人/大洋洲。在AIM中 3，我建议量化不同人类基因组中“幽灵”变异的选择景观， 祖先基因流来自现已灭绝的“幽灵”种群。为达致这个目标，我们会集中于（a）改善 MigSelect程序，用于量化IM下来自“幽灵”种群的基因流导致的连锁选择效应 模型，和（B）一个更大的，更广泛的研究功能基因组变异在不同的人类 人口，包括来自非洲的高质量基因组，补充更完整的尼安德特人， 其他非人类的古人类基因组，这将有助于我们描绘人类进化历史的模式， 了解古代基因流的功能后果。这些发现也有直接的后果 了解现代人类祖先和疾病等位基因的进化。重要的是，这款R15将训练许多 基因组学和生物信息学的本科生和研究生， 生物医学和数据科学。