A Desktop Assembly and Analysis Pipeline for Next-gen Metagenomic Sequencing

描述（由申请人提供）：在过去的十年中，细菌和古细菌群落已经在几乎每一个被检查的生态位中被确定，从酸性矿井排水到对流层云再到人体。了解这些群落的组成和相互作用对于了解每个生态系统的功能至关重要。宏基因组学（或群落基因组学）是一种方法，通过这种方法可以确定自然存在的微生物群落的集体基因组内容，而无需独立分离和培养其组成部分。组装后，可以获得群落的遗传框架，包括种群结构、系统发育多样性以及新的遗传和生化活动的关键信息。这些发现对生物技术、医学和生态学的潜在影响是巨大的。最近，新一代测序技术（如罗氏/454，Illumina和Life technologies (SOLiD)）已经取代了传统的Sanger测序，用于宏基因组数据生成。这些都是低成本、无克隆、大规模并行的技术，能够在一台机器运行中产生多达250千兆字节的数据。这种测序水平使得重建低丰度基因组以及从群落样本中确定种内异质性成为可能。然而，下一代技术也提出了自己的计算挑战，包括需要处理的数据量以及每种技术特有的不同读取长度、错误模型和格式。这些复杂性加上社区本身的复杂性，使得研究人员不得不拼凑各种软件工具组合来处理和分析他们的数据。大多数能够处理这些大型复杂数据集的软件也需要大量的计算资源和计算机专业知识，超出了通常配备的实验室的能力。这些困难继续对宏基因组学提供的科学和技术进步产生严重的抑制作用。该提案的长期目标是开发一个无缝的，商业级的宏基因组序列组装和分析管道，可以完全扩展到任何规模的项目。拟议中的软件将易于使用，并在台式计算机上运行，成本低于5000美元，因此任何合理资助的实验室或诊所都可以利用宏基因组技术。为了实现这一目标，第一阶段的提案侧重于在台式计算机上处理大量下一代宏基因组数据集的中心任务的解决方案。我们将通过三个关键步骤来评估我们新的非内存绑定组装引擎XNG是否能够应对挑战：1)去除来自污染宿主DNA的reads， 2)根据与本地参考基因组数据库的匹配，将潜在的数亿个剩余reads“招募”到适当的系统发育箱中，3)将给定物种的多个菌株的基因组序列转换为单个注释条目（“泛基因组”），以增强reads招募和下游注释。

DESCRIPTION (provided by applicant): Over the last decade, bacterial and archaeal communities have been identified in virtually every ecological niche examined, from acid mine drainage to tropospheric clouds to the human body. Understanding the make-up and interactions within these communities is crucial in understanding how each ecosystem functions. Metagenomics (or community genomics) is the approach whereby the collective genomic content of a naturally occurring microbial community can be determined without the need to isolate and culture its constituents independently. When assembled, the genetic framework of the community, including critical information on population structure, phylogenetic diversity, as well as novel genetic and biochemical activities can be obtained. The potential impact of such findings on biotechnology, medicine and ecology are enormous. Recently, next-gen sequencing technologies (e.g. Roche/454, Illumina, and Life Technologies (SOLiD)) have replaced traditional Sanger sequencing for metagenomic data generation. These are cost effective, clone-free, massively parallel technologies capable of producing as many as 250 gigabases of data in a single machine run. That level of sequencing makes it feasible to reconstruct even low abundance genomes as well as determine intraspecies heterogeneity from a community sample. However, the next-gen technologies also present their own computational challenges including the sheer volume of data to be processed together with the different read lengths, error models and formats unique to each technology. These complexities together with those posed the communities themselves together has left researchers to cobble together various combinations of software tools to process and analyze their data. Most software that can handle these large, complex data sets also require substantial computing resources and computer expertise beyond that of a normally equipped lab. These difficulties continue to have a serious stifling effect on the advances in science and technology that metagenomics offers. The long term goal of this proposal is to develop a seamless, commercial-grade metagenomic sequence assembly and analysis pipeline that is fully scalable to any size project. The proposed software will be easy to use and run on a desktop computer costing less than $5000 so that any reasonably funded laboratory or clinic can exploit metagenomic technology. Toward that goal, this Phase I proposal focuses on a solution to the central task of processing massive next-gen metagenomic data sets on a desktop computer. We will evaluate whether our new non-memory bound assembly engine, XNG, can meet the challenges in three crucial steps: 1) removing reads derived from contaminating host DNA, 2) "recruiting" the potentially hundreds of millions remaining reads into appropriate phylogenetic bins based on matches to a local reference genome database, and 3) converting genome sequences from multiple strains of a given species into a single annotated entry (the "pan-genome") for enhanced read recruitment and downstream annotation. PUBLIC HEALTH RELEVANCE: Human health and medicine are greatly influenced by the microbial communities that are integral parts of our bodies as well as those that produce useful antibiotics for example. To better exploit the potential of these communities, metagenomic studies are producing vast amounts of Next-gen DNA sequence data to decipher their collective genetic content. This project focuses on developing computer software capable of reconstructing microbial community genomes and analyzing their content.