ABI Innovation: A new computational framework for the prediction of microbiome dynamics

ABI Innovation：用于预测微生物组动态的新计算框架

• 批准号: 1458347
• 负责人: Vanni Bucci
• 金额: $ 49.17万
• 依托单位: University of Massachusetts, Dartmouth
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2015
• 资助国家: 美国
• 起止时间: 2015-09-01 至 2019-08-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1458347&HistoricalAwards=false
• 关键词: ABI; Innovation; new; computational; framework

The dynamics of microbial communities play a fundamental role in the functioning of many natural, engineered and host-associated systems. Even though the application of DNA sequencing technologies has allowed profiling the response of these communities to external perturbations, the important knowledge resulting from this approach stems from descriptive and correlation-based analysis of these data. This strongly limits the understanding of the ecology (e.g. how the microbes interact) of these systems and, more importantly, hinders the ability to make quantitative predictions. This project will deliver new theoretical methods and related computational algorithms that, for the first time, allow forecasting microbiome dynamics that are typically constrained with sequencing surveys. This will benefit researchers working on host-associated and environmental microbiomes as it will enable them to computationally explore scenarios that are difficult to set-up experimentally. The tools developed in this project will be delivered as an open-source, freely downloadable and upgradable package, and will encourage and enable end-user contributions with the novel scripts and algorithms provided. Multiple graduate and undergraduate students will be included in the project and will benefit from interdisciplinary hands-on training in mathematical and computational biology, statistics, and microbial genetics. As the proposed methods combine concept from multi-linear regression and solution of large systems of differential equations, they will perfectly integrate with coursework in Biostatistics and Theoretical Biology at UMass Dartmouth. This research will deliver the first computational suite that allows for simulating and predicting microbiome dynamics consistent with metagenomics observations. This will be achieved by: 1) the development of new time-reverse engineering inference methods solved by a combination of regularized regression and quadratic programming for the estimation of an optimal set of model parameters, and 2) the application of computational tools for metagenome reconstruction based on modeling predictions. This research will also include the development of explicit and implicit numerical methods for the solution of large systems of differential equations to predict microbiome transient dynamics and the use of linear stability analysis to determine all possible microbiome predicted configuration states in response to different sets of perturbations. Method testing and validation against data from simple in silico and in vitro microbial ecosystem of known ecological structure will allow accuracy testing of the proposed approaches both for predicting temporal dynamics and in recovering the correct microbial interaction network in response to external perturbation. Method application (and validation) on diverse datasets will allow testing of fundamental hypotheses about the role of the intestinal microbiome in resisting colonization by foreign bacteria and in shaping host immunity. More information about this project can be found at: http://www.vannibucci.org/research-interests.html

微生物群落的动态在许多自然、工程和宿主相关系统的功能中起着重要作用。尽管DNA测序技术的应用已经允许对这些群落对外部扰动的反应进行分析，但这种方法所产生的重要知识源于对这些数据的描述性和相关性分析。这极大地限制了对这些系统的生态学（例如微生物如何相互作用）的理解，更重要的是，阻碍了进行定量预测的能力。该项目将提供新的理论方法和相关的计算算法，首次允许预测通常受测序调查限制的微生物组动态。这将有利于研究宿主相关和环境微生物组的研究人员，因为这将使他们能够通过计算探索难以通过实验建立的场景。本项目开发的工具将作为开源、可免费下载和升级的软件包交付，并将鼓励和支持终端用户对所提供的新颖脚本和算法做出贡献。多名研究生和本科生将参与该项目，并将受益于数学和计算生物学、统计学和微生物遗传学的跨学科实践培训。由于所提出的方法结合了多元线性回归的概念和大型微分方程组的解，它们将完美地与麻省大学达特茅斯分校的生物统计学和理论生物学课程相结合。这项研究将提供第一个计算套件，允许模拟和预测与宏基因组学观察一致的微生物组动力学。这将通过：1)发展新的时间逆向工程推理方法，通过正则化回归和二次规划的组合来解决最优模型参数集的估计，以及2)应用基于建模预测的宏基因组重建计算工具来实现。这项研究还将包括开发显式和隐式数值方法来解决大型微分方程系统，以预测微生物组的瞬态动力学，并使用线性稳定性分析来确定响应不同扰动集的所有可能的微生物组预测配置状态。针对已知生态结构的简单的室内和体外微生物生态系统的数据进行方法测试和验证，将允许对所提出的方法进行准确性测试，以预测时间动态和恢复响应外部扰动的正确微生物相互作用网络。对不同数据集的方法应用（和验证）将允许测试关于肠道微生物群在抵抗外来细菌定植和形成宿主免疫中的作用的基本假设。关于这个项目的更多信息可以在http://www.vannibucci.org/research-interests.html上找到