III: Small: Computational Inference of Microbial Community Structures from Environmental Shotgun Reads

微生物在我们的日常生活中扮演着重要的角色。然而，大多数微生物物种不容易被研究，因为它们不能与它们的环境分离，即微生物群落。基于直接从微生物群落中提取的混合DNA序列，宏基因组学已成为研究微生物和微生物群落的一个重要领域。“群落结构鉴定”是宏基因组学的核心问题之一。目的是从提取的混合DNA序列中确定存在于特定微生物群落中的物种及其相对丰度。由于公共数据库中的可培养物种和下一代测序平台的短DNA序列的分类学偏差，迫切需要创建能够有效解决“群落结构识别”问题的计算方法。为了最大限度地减少分类偏差和DNA序列短长度的影响，拟议的研究将创建一个基于k元组频率的统计框架，以筛选可能来自同一物种的reads。在此框架下，将设计一系列算法，通过整合大规模基因组序列及其注释来推断群落结构。研究活动将根据模拟和实验数据进行评估。一个配套的软件包也将开发并免费发布给研究社区。所提出的方法和工具将有助于为进一步研究微生物奠定基础，从而显著推进对微生物和微生物群落的科学认识。本项目将为少数民族高中生和大学生提供研究经验。它还将通过课程开发、研讨会、指导活动和年度研究专题讨论会对本科生和研究生进行宏基因组学教育。此外，该项目将通过出版物、会议演讲、自由软件开发等方式传播宏基因组学的研究成果。最后，通过解决宏基因组学的核心问题并提供各种准确的推断，本研究将极大地推进当前对微生物和微生物群落的认识。

Microbes play important roles in our everyday life. However, the majority of microbial species cannot be readily studied because they cannot be separated from their environments, i.e., microbial communities. Based on mixed DNA sequences directly extracted from microbial communities, metagenomics has emerged as an important field studying microbes and microbial communities. "Community structure identification" is one of the core problems in metagenomics. The goal is to identify species present in a specific microbial community and their relative abundance from the extracted mixed DNA sequences. Because of the taxanomic bias towards culturable species in public databases and short DNA sequences from next generation sequencing platforms, there is an urgent need to create computational methods that can effectively address the "community structure identification" problem. In order to minimize the effect of the taxonomic bias and the short length of DNA sequences, the proposed research will create a statistical framework to bin reads that are likely from the same species based on k-tuple frequencies. With this framework, a series of algorithms will be designed to infer the community structures by integrating large-scale genomic sequences and their annotations. The research activities will be evaluated based on both simulated and experimental data. An accompanying software package will also be developed and released to the research community for free. The proposed methods and tools will help lay the foundation for further studying the microbes towards significantly advancing the scientific understanding of microbes and microbial communities.This project will provide research experience for minority high school students and undergraduates. It will also educate undergraduates and graduates in metagenomics through curriculum development, seminars, mentoring activity, and annual research symposiums. In addition, this project will disseminate the research results on metagenomics through publications, conference presentations, free software development, and others. Finally, by attacking a core problem in metagenomics and providing a variety of accurate inferences, this research will greatly advance the current knowledge of microbes and microbial communities.