Developing computational, statistical and machine learning methods to uncover biological mechanisms of complex phenotypes

开发计算、统计和机器学习方法来揭示复杂表型的生物学机制

• 批准号: RGPIN-2021-04062
• 负责人: Shooshtari, Parisa
• 金额: $ 1.75万
• 依托单位: University of Western Ontario
• 依托单位国家: 加拿大
• 项目类别: Discovery Grants Program - Individual
• 财政年份: 2021
• 资助国家: 加拿大
• 起止时间: 2021-01-01 至 2022-12-31
• 项目状态: 已结题
• 来源: https://www.nserc-crsng.gc.ca/ase-oro/Details-Detailles_eng.asp?id=742820
• 关键词: Developing; computational; statistical; machine; learning

My research program develops computational methods to analyze large omic datasets to understand gene regulatory mechanisms of complex human traits. My long-term objective is to understand how the epigenetic landscape affects the genetic background in a complex, multifactorial fashion. Integration of approaches including genome-wide association studies and single-cell sequencing technology is critical in understanding complex traits in multicellular organisms. Obesity is a complex trait defined as an excessive or abnormal accumulation of fat, and it can increase the risk of various health problems. The prevalence rate of obesity is increasing in Canada, imposing significant health, economical, and social impacts. A lack of understanding of molecular mechanisms underlying obesity prevents developing preventative approaches. As my short-term objective, I will focus on studying cellular and molecular mechanisms of BMI, which is the most commonly used proxy for obesity. Several studies show that BMI-associated variants are enriched on regulatory sites of certain cell types, supporting a role in gene regulation in BMI. However, the specific gene regulatory mechanisms underlying BMI are still largely unknown. To gain mechanistic insight into gene regulation in BMI, I will develop effective computational methods to: Aim 1: Identify specific BMI-relevant regulatory factors. Transcription factors (TF) play a central role in gene regulation of human traits. However, the specific TFs relevant to BMI and the molecular mechanisms underlying their effects are largely unknown. A major thrust of my research will develop computational approaches to uncover BMI-relevant TFs, and understand functional mechanisms through which the associated variants affect regulatory activities of their target TFs. Aim 2: Uncover specific BMI-relevant cell types. Identifying specific cells in which regulation of obesity-relevant genes are affected is crucial for uncovering gene regulatory mechanisms. Multiple studies using bulk epigenomic data support the role of brain cells in BMI. However, brain cell types are highly heterogeneous and the specific subsets of brain cells that are most relevant cannot be identified using bulk sequencing data. I will develop a computational method that integrates single-cell epigenetic and genetics association data to identify BMI-relevant cells. This proposal outlines the beginnings of a long-term NSERC-funded research program uniquely aimed at generating a fundamental understanding of gene regulatory mechanisms in complex traits. While initially using BMI as the target complex trait, this work will develop a methodology that can be applied to the myriad of other complex traits that exist. The work will vastly enhance our knowledge of trait-relevant regulatory pathways and their sites of actions (ie. cell types). I predict our methods will become widely used resources by the genomics community, enabling discoveries in many complex traits.

我的研究项目开发了计算方法来分析大型基因组数据集，以了解复杂人类特征的基因调控机制。我的长期目标是了解表观遗传景观如何以复杂的、多因素的方式影响遗传背景。整合包括全基因组关联研究和单细胞测序技术在内的方法对于理解多细胞生物的复杂性状至关重要。肥胖是一种复杂的特征，被定义为脂肪的过度或异常积累，它可以增加各种健康问题的风险。加拿大的肥胖患病率正在上升，对健康、经济和社会造成了重大影响。缺乏对肥胖分子机制的理解阻碍了预防方法的发展。作为我的短期目标，我将重点研究BMI的细胞和分子机制，BMI是最常用的肥胖指标。一些研究表明，BMI相关变异在某些细胞类型的调控位点富集，支持BMI基因调控的作用。然而，BMI的具体基因调控机制在很大程度上仍是未知的。为了深入了解BMI基因调控的机制，我将开发有效的计算方法来：目标1：确定特定的BMI相关调控因子。转录因子（TF）在人类性状的基因调控中起着核心作用。然而，与BMI相关的特定tf及其影响的分子机制在很大程度上是未知的。我研究的一个主要方向是开发计算方法来揭示bmi相关的tf，并了解相关变异影响其目标tf调控活动的功能机制。目标2：揭示特定的bmi相关细胞类型。鉴定受肥胖相关基因调控影响的特定细胞对于揭示基因调控机制至关重要。使用大量表观基因组数据的多项研究支持脑细胞在BMI中的作用。然而，脑细胞类型是高度异质性的，并且最相关的特定脑细胞亚群无法使用批量测序数据进行鉴定。我将开发一种整合单细胞表观遗传学和遗传学关联数据的计算方法来识别bmi相关细胞。该提案概述了nserc资助的一项长期研究计划的开端，该计划旨在对复杂性状中的基因调控机制产生基本的理解。虽然最初使用BMI作为目标复杂特征，但这项工作将开发一种可以应用于无数其他存在的复杂特征的方法。这项工作将极大地提高我们对性状相关的调控途径及其作用位点的认识。细胞类型)。我预测我们的方法将成为基因组学社区广泛使用的资源，使许多复杂性状的发现成为可能。