DMS/NIGMS 1: Modeling Microbial Community Response to Invasion: A Multi-Omics and Multifacton

大多数微生物生活在包含数百或数千个物种的群落中，每个物种都参与其中。 in a rich丰富web网of interactions相互作用.这些相互作用的复杂性使得定量预测 社区动力困难。为了克服这一点，自然群落的简单代理，指定 "模式微生物群落"，旨在支持实验室研究。这些模型 复杂到足以展示社区特有的现象，但又简单到足以揭示治理 社区互动的原则。入侵是微生物最不稳定的事件之一 社区可能会经历，往往导致社区功能障碍或宿主疾病。我们建议使用 THOR，我们开发的一个模型社区，用于描述假单胞菌入侵的反应 铜绿。我们将跟踪种群动态功能剖面分子相互作用。 这项研究的多组学和多因素性质为统计提供了多方面的机会。 创新一个真正的综合分析不能简单地进行平行的假设检验， 需要一种将数据源混合到统一的分子测试框架中的差异测试框架 互动网络我们将借鉴选择性推理和多组学网络分析的进展， 开发阐明推动社区反应的分子相互作用的方法。这将使我们 定制干预措施，塑造THOR模型微生物群落的动态。我们建议： 1.目的1：使用代谢组学功能性描述THOR对铜绿假单胞菌的社区反应， 元转录组学和16S rRNA测序，并建立相关的数据管理工作流程。这些 是核心的数据生成实验和质量控制步骤， 补充意见的雷神之锤下入侵。 2.目标2：开发对相互作用效应敏感的差异测试方法， 模块级错误发现率。我们将介绍微分的选择性推理方法， 相互作用体模块，如激活的生物合成途径。 3.目标3：整合用于差异相互作用组分析的软件并通过实验验证敲除 目标的这些实验将说明一种数据驱动的方法来控制入侵动态，我们的 软件将使这种分析更容易获得。

Most microorganisms live in communities containing hundreds or thousands of species, each engaging in a rich web of interactions. The complexity of these interactions makes quantitative predictions about community dynamics difficult. To overcome this, simple proxies for natural communities, designated "model microbial communities," have been designed to support laboratory study. These models are complex enough to exhibit community-specific phenomena but simple enough to reveal governing principles of community interaction. Invasions are among the most destabilizing events that a microbial community can experience, often resulting in community dysfunction or host disease. We propose to use THOR, a model community that we developed, to characterize the response to invasion by Pseudomonas aeruginosa. We will track population dynamics functionally profile molecular interactions. The multi-omics and multifactorial nature of this study present multifaceted opportunities for statistical innovation. A truly integrative analysis cannot simply perform parallel hypothesis tests across assays, and there is a need for a differential testing framework that blends data sources into a unified molecular interaction network. We will draw from advances in selective inference and multi-omics network analysis to develop methods that illuminate the molecular interactions driving community response. This will allow us to tailor interventions that shape dynamics in the THOR model microbial community. We propose: 1. Aim 1: Functionally profile THOR's community response to P. aeruginosa using metabolomics, metatranscriptomics, and 16S rRNA sequencing and establish associated data curation workflows. These are the core data-generating experiments and quality control steps that provide accurate and complementary views of THOR under invasion. 2. Aim 2: Develop differential testing methods that are sensitive to interaction effects and that control module-level false discovery rates. We will introduce methods for selective inference of differential interactome modules, like activated biosynthetic pathways. 3. Aim 3: Consolidate software for differential interactome analysis and experimentally validate knockout targets. These experiments will illustrate a data-driven approach to control invasion dynamics, and our software will make such analysis easily accessible.