Machine learning-based multi-omics modeling and CRISPR/Cas9-mediated gene editing in elucidating molecular transducer of physical activity

摘要 定期锻炼（体育活动）是促进健康和对抗非健康疾病的最有效干预措施。 传染病（NCD）。然而，我们对这些分子的理解， 运动的好处是模糊的。美国国立卫生研究院共同基金项目“身体活动的分子传感器 Consortium（MoTrPAC）是一项大规模的发现研究，旨在了解 运动训练，它已经发布了第一批多组学数据，包括RNA-seq， 亚硫酸氢盐测序、蛋白质组学、磷酸化蛋白质组学、乙酰蛋白质组学和靶向和 非靶向代谢组学，来自急性发作后大鼠在不同时间点收集的5种组织。 耐力训练这些工作为阐明分子转导子的功能奠定了基础。 体力活动。我们最近在四个领域取得了重大进展，使我们能够探讨这些问题。 以前所未有的方式收集数据并阐明机制。具体而言，1）我们已经获得了类似的 急性和长期耐力运动后小鼠4种组织中的时程、转录组学数据， 开发了机器学习能力，用于挖掘多组学数据，以识别 2）我们已经完善了CRISPR/Cas9介导的基因编辑，以产生损失- 功能缺失敲入小鼠以及产生组织特异性、可诱导的功能获得性转基因小鼠的技术 小鼠; 3）我们已经建立了全面的小鼠表型分析;和4）我们已经成功地 在阐明细胞外超氧化物歧化酶（EcSOD）的调节和功能方面的经验， 在骨骼肌中表达并通过耐力运动促进的因子，介导健康益处， 预防疾病。我们假设耐力运动促进了一个基因的表达和释放， 一个或多个组织/器官的一种或多种体液因子，这是充分和必要的介导的， 锻炼的健康益处。为此，我们提议 1)基于机器学习的多组学建模识别身体活动的候选分子转换器。 2)使用CRISPR/Cas9- 3产生功能缺失敲入和组织特异性功能获得转基因小鼠 介导的基因编辑和转基因。 3)阐明身体活动的候选分子转换器在运动的健康益处中的作用。 实验设计和模型系统在概念和技术上都是创新的。其成果将 显著提高对具有巨大潜在影响的运动诱导适应的机制理解 关于NCD治疗方法的未来发展。

ABSTRACT Regular exercise (physical activity) is the most effective intervention that promotes health and combats non- communicable disease (NCD). However, our understanding of the molecule(s) responsible for the superb benefits of exercise is obscure. The NIH Common Fund project “Molecular Transducers of Physical Activity Consortium (MoTrPAC)” is a large-scale discovery study designed to understand the molecular responses to exercise training, which has released the first batch of multi-omics data, including RNA-seq, Reduced Representation Bisulfite Sequencing, proteomics, phosphoproteomics, acetylproteomics, and targeted and untargeted metabolomics, from 5 tissues collected at different time points in rats following an acute bout of endurance exercise. These endeavors have laid a solid foundation for elucidation of the molecular transducer of physical activity. We have recently made significant progress in four areas, which poised us to explore these data and elucidate the mechanism(s) in an unprecedented manner. Specifically, 1) We have obtained similar time-course, transcriptomics data in 4 tissues in mice following acute and long-term endurance exercise and developed machine learning capability for mining the multi-omics data for identification of regulatory factors that mediate the exercise benefits; 2) We have perfected CRISPR/Cas9-mediated gene editing for generation of loss- of-function knock-in mice as well as techniques to generate tissue-specific, inducible gain-of-function transgenic mice; 3) We have established comprehensive phenotypic analysis in mice; and 4) We have had a successful experience in elucidating the regulation and function of extracellular superoxide dismutase (EcSOD), a humoral factor expressed in skeletal muscle and promoted by endurance exercise, in mediating the health benefits and protection against diseases. We hypothesize that endurance exercise promotes expression and release of one or more humoral factors from one or multiple tissues/organs, which is sufficient and necessary mediating the health benefits of exercise. To this end, we propose 1) Identify candidate molecular transducers of physical activity by machine learning-based multi-omics modeling. 2) Generate loss-of-function knock-in and tissue-specific, gain-of-function transgenic mice using CRISPR/Cas9- mediated gene editing and transgenesis. 3) Elucidate the role of the candidate molecular transducers of physical activity in health benefits of exercise. The experimental design and model systems are both conceptually and technically innovative. The findings will significantly improve the mechanistic understanding of exercise-induced adaptations with great potential impact on the future development of therapeutics for NCD.