Dominance on the Human Genome and Non-additive Polygenic Models for Predicting Complex Traits

人类基因组的主导地位和用于预测复杂性状的非加性多基因模型

• 批准号: 10755393
• 负责人: Xinjun Zhang
• 金额: $ 24.9万
• 依托单位: UNIVERSITY OF MICHIGAN AT ANN ARBOR
• 依托单位国家: 美国
• 财政年份: 2021
• 资助国家: 美国
• 起止时间: 2021-08-01 至 2026-01-31
• 项目状态: 未结题
• 来源: https://reporter.nih.gov/project-details/10755393
• 关键词: Accounting; Address; Admixture; Affect; Biological; Biology; Collaborations; Communities; Complex; Computing Methodologies; Data Analyses; Demography; Development; Disease; Ecology; European; Evolution; Genetic; Genome; Genomic Segment; Genomics; Human; Human Genome; Individual; Journals; Knowledge; Machine Learning; Mentors; Methods; Minority Groups; Modeling; Modernization; Mutation; Outcome; Pattern; Phase; Phenotype; Play; Population; Population Genetics; Population Heterogeneity; Recording of previous events; Reporting; Research Personnel; Research Training; Resolution; Role; Series; Shapes; Statistical Methods; Study Subject; Testing; Training; Variant; Work; biobank; disorder risk; experience; fitness; functional genomics; genetic variant; genome wide association study; genome-wide; genomic data; human population genetics; improved; insight; machine learning method; novel; polygenic risk score; precision medicine; predictive modeling; risk prediction; simulation; training opportunity; trait

Project Abstract Dominance is one of the most fundamental concepts in genetics and has many key implications in population genetics, as it ultimately determines how selection manifests in a population. However, despite its unarguable importance, dominance is also one of the least characterized quantities in genetics, especially in humans, with the major challenge being current methods cannot distinguish dominance from the fitness effect of genomic variants. This proposed K99/R00 work will systematically address this longstanding problem from a dual- perspectives, by inferring dominance in humans and quantitatively model its role in shaping the phenotypes of complex traits and diseases. Specifically, in Aim1, I will develop a powerful machine learning-based method to infer the realistic distribution of dominance on the human genome in megabase-scale, leveraging archaic introgressed ancestry in non-African populations that is sensitive to dominance variation in genomic regions. In Aim 2, I will develop non-additive polygenic models accounting for dominance in full genomic regions to identify complex traits profiled in UK Biobank that deviate from additive models, improve the accuracy of phenotype and disease risk predictions, and contribute to an in-depth understanding of complex trait biology. Finally, in Aim 3 (R00 phase), I will extend these approaches to infer dominance variation in worldwide populations and investigate how dominance, combined with selection and admixture, determines complex trait phenotypes in diverse human populations. The mentored phase of this work will take place at the Department of Ecology and Evolutionary Biology at UCLA, where Dr. Zhang will have access to rich training opportunities and be supported by active scientific communities, including numerous seminar series, journal clubs, and networking activities. Dr. Kirk Lohmueller (primary mentor) and Dr. Sriram Sankararaman (co-mentor) will train Dr. Zhang in computational and statistical methods in population genetics, machine learning applications, and large-scale disease association data analysis. The research trainings, collaborations, and professional development during the K99 phase will assist Dr. Zhang in becoming an independent investigator in human population genetics.

项目摘要 优势是遗传学中最基本的概念之一，对人口有许多关键影响 遗传学，因为它最终决定了选择在群体中的表现方式。然而，尽管其无可争议 重要性，优势也是遗传学中最不具有特征的数量之一，尤其是在人类中， 主要挑战是当前的方法无法区分基因组的优势和适应度效应 变体。这项拟议的 K99/R00 工作将从双方面系统地解决这个长期存在的问题 通过推断人类的主导地位并定量模拟其在塑造人类表型中的作用 复杂的性状和疾病。具体来说，在 Aim1 中，我将开发一种强大的基于机器学习的方法来 利用古老的算法，推断出百万碱基规模的人类基因组优势的实际分布 非非洲人群中的渐渗祖先对基因组区域的显性变异敏感。在 目标 2，我将开发非加性多基因模型来解释全基因组区域的优势，以识别 英国生物库中描述的复杂性状偏离了加性模型，提高了表型和 疾病风险预测，并有助于深入了解复杂的性状生物学。最后，在目标 3 中 （R00 阶段），我将扩展这些方法来推断全球人口的优势变化， 研究显性如何与选择和混合相结合，决定复杂的性状表型 多样化的人口。这项工作的指导阶段将在生态学系进行 加州大学洛杉矶分校进化生物学，张博士将获得丰富的培训机会并得到支持 活跃的科学界，包括众多研讨会系列、期刊俱乐部和网络活动。博士。 Kirk Lohmueller（主要导师）和 Sriram Sankararaman 博士（联合导师）将对张博士进行计算方面的培训 群体遗传学、机器学习应用和大规模疾病中的统计方法 关联数据分析。 K99 期间的研究培训、合作和专业发展 该阶段将帮助张博士成为人类群体遗传学的独立研究者。