CAREER: Computational tools for analyzing and interpreting DNA methylation

该奖项的全部或部分资金来自《2021年美国救援计划法案》(公法117-2)。DNA甲基化是一个关键的生物过程，在基因调控中发挥着如此重要的作用，以至于它经常被称为“DNA的第五个碱基”。除了在发育和整个衰老过程中发挥重要作用外，DNA甲基化异常还被发现会导致疾病，包括癌症和神经疾病。然而，关于DNA甲基化变化发生在哪里以及它们影响了哪些基因，仍有许多未知之处。这一点，再加上它的动态性质，可以在身体的不同区域，在整个生命周期，并对环境的反应，使准确地确定特定甲基化模式与感兴趣的生物现象的联系，并解释其下游影响具有挑战性。该项目将产生一套新的工具，以实现跨实验技术的甲基化位点分析，在不同组织和细胞类型的背景下分析甲基化数据，并预测在不同条件下变化的甲基化位点的下游影响。这些工具将使研究人员能够根据现有知识更好地解释他们的甲基化数据，这反过来可以导致对基本生物过程(例如，发育、衰老)的更好理解。除了将所有方法作为开放源码软件提供外，还将开发预测数据库和交互式可视化，并在网上提供。通过这样做，没有编程经验的生物研究人员可以使用基于查询的系统来轻松地在他们自己的数据背景下做出和探索预测。该项目还有一个本地推广部分，通过让高中生物学教师和社区大学生参与新设计的数据驱动课程和研究机会，帮助增加STEM中未被充分代表的少数族裔的多样性和持久性。研究将采用其他领域用于归因的尖端深度学习方法，以增加DNA甲基化平台的覆盖率，该平台描述了10%的所有位点，节省了时间和资源；使用反映自然组织和细胞类型依赖的分层框架来寻找特定位置的甲基化标志；并结合已知的蛋白质-蛋白质相互作用，将CpG位点与生物功能联系起来，而不是现有的主要捕获近端调控关系的方法。除了这项工作的方法学贡献外，该项目本身就是跨学科的，将为理解基本的生物学问题奠定基础，例如甲基化如何调节和维持组织特异性。这个项目的工具和结果可以在：https://cs.rice.edu/~vy/.This奖项反映了国家科学基金会的法定使命，并通过使用基金会的智力价值和更广泛的影响审查标准进行评估，被认为值得支持。

This award is funded in whole or in part under the American Rescue Plan Act of 2021 (Public Law 117-2).DNA methylation is a critical biological process that plays such an important role in gene regulation that it has often been referred to as the “fifth base of DNA.” In addition to playing an important role during development and throughout aging, aberrations in DNA methylation have been discovered to cause disease, including cancer and neurological disorders. Yet much is still unknown regarding where DNA methylation changes occur and which genes they impact. This, coupled with its dynamic nature, which can vary across regions of the body, throughout the life cycle, and in response to the environment, makes accurately identifying the association of specific methylation patterns to biological phenomena of interest and interpreting their downstream impacts challenging. The project will result in a suite of new tools to enable analysis of methylation sites across experimental technologies, analyze methylation data in the context of different tissues and cell types, and predict the downstream impacts of methylation sites that are changing across conditions. These tools will enable researchers to better interpret their methylation data in light of existing knowledge, which can in turn result in improved understanding of fundamental biological processes (e.g., development, aging). In addition to making all methods available as open-source software, databases of predictions and interactive visualizations will be developed and accessible online. By doing so, biological researchers with no programming experience can use a query-based system to easily make and explore predictions in the context of their own data. This project also has a local outreach component to help increase the diversity and persistence of underrepresented minorities in STEM by engaging high school biology teachers and community college students with newly designed data driven curricula as well as research opportunities.The research will adapt cutting edge deep learning methods used for imputation in other domains to increase the coverage of DNA methylation platforms that profile 10% of all sites, saving time and resources; use a hierarchical framework that mirrors natural tissue and cell type dependencies to find location-specific methylation hallmarks; and incorporate known protein-protein interactions to associate CpG sites with biological function, beyond existing methods which primarily capture proximal regulatory relationships. In addition to the methodological contributions of this work, the project is inherently interdisciplinary and will lay the groundwork for understanding fundamental biological questions such as how methylation regulates and maintains tissue specificity. The tools and results from this project can be found at: https://cs.rice.edu/~vy/.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.