Statistical Methods for Enhanced Mapping of Microbiome Relationships

项目摘要/摘要 人类微生物群与广泛的人类疾病密切相关，并代表着一种势在必行的 减轻这些疾病负担的途径，特别是在微生物组明显可改变的情况下。 这最终导致了以微生物为导向的风险降低和临床干预，这已被证明 在从感染到癌症免疫治疗的各种情况下都有效。然而，尽管有一些高- 除了个人资料的成功，许多其他研究都失败了。这些失败背后的一个中心主题，甚至是许多 成功的故事，是我们对微生物如何相互作用的基本有限的理解， 宿主基因组学，并与结果。最近，评估和绘制这些关系的努力正在进行 在大规模的档案研究中，但不幸的是，用于阐明这些联系的工具可能是 由于强大的基础，功能不足、难以解释，甚至会出现严重的误报 假设。因此，在问题的驱使下，三个最大和最丰富的微生物组图谱 在这项研究中，这项提案试图通过处理四个主要领域来填补方法学文献中的关键空白。 具体地说，我们的目标是开发一套全面的工具，用于(1)增强的微生物共生网络 构建；(2)加强发现与单个微生物分类群相关的SNPs和稀有变异；以及 (3)用孟德尔随机化方法评价微生物的作用。这些方法都是基于 严格的先前数据强调了问题的重要性以及现有战略的局限性。 我们的工作是由三个最大和最丰富的微生物组图谱的分析推动的，并将直接使分析成为可能 研究，包括研究肠道微生物组的MEC和SOL队列，以及探索 怀孕期间的阴道微生物群。因此，我们开发的方法有可能改进我们的 微生物组的基本知识，并推动该领域朝着加强风险降低和临床 干预措施。

PROJECT SUMMARY/ABSTRACT The human microbiome is integrally related to a vast range of human disorders and represents an imperative gateway toward mitigating the burden of these diseases, particularly as the microbiome is eminently modifiable. This has culminated in microbially oriented risk reduction and clinical interventions, which have proven efficacious in diverse situations ranging from infections to cancer immunotherapy. However, despite some high- profile successes, many other studies have failed. A central theme underlying these failures, and even many of the success stories, is our fundamentally limited understanding of how microbes interact with each other, with host genomics, and with outcomes. Recently, efforts to assess and map these relationships are taking place within large-scale profiling studies, but unfortunately, the tools used for elucidating these connections may be underpowered, difficult to interpret, or even subject to severe false positives due to strong underlying assumptions. Therefore, motivated by problems within three of the largest and richest microbiome profiling studies, this proposal seeks to fill critical gaps in the methodological literature by addressing four major areas. Specifically, we aim to develop a comprehensive suite of tools for (1) enhanced microbial co-occurrence network construction; (2) enhanced discovery of SNPs and rare variants associated with individual microbial taxa; and (3) assessing the role of microbes through Mendelian Randomization. These approaches are all based on rigorous prior data emphasizing the importance of the problems as well as the limitations of existing strategies. Our work is motivated by and will directly enable analyses in three of the largest and richest microbiome profiling studies, including the MEC and SOL cohorts which study the gut microbiome, and the PIN cohort, which explores the vaginal microbiome in pregnancy. Accordingly, the methods we develop have the potential to improve our fundamental knowledge of the microbiome and propel the field towards enhanced risk reduction and clinical interventions.