ABI Development: A Novel Protein Fragment Assembler for Metagenomic Data Analysis

该奖项授予J. Craig Venter研究所（JCVI），旨在开发分析宏基因组序列数据的新方法。宏基因组学是指利用独立于培养的技术研究微生物群落的基因组内容。这种模式彻底改变了微生物学领域，特别是考虑到我们无法培养存在于许多环境中的大多数微生物。新一代测序技术通常用于从宏基因组样本中生成大量核苷酸序列读取数据。从这些数据中鉴定出全长蛋白质序列，可以对组成微生物的代谢潜力进行全面而准确的分析。这通常是通过组装核苷酸序列reads，然后使用生成的contigs集作为蛋白质鉴定的底物来实现的。然而，宏基因组组装通常是非常碎片化的，产生较短的contigs，并且也留下了很大一部分未组装的reads，从而限制了该方法的实用性。该项目将开发一个序列组装框架，用于直接从核苷酸读取上鉴定的短肽片段重建全长蛋白质序列。这种方法的动机是两个观察？(a)在原核生物基因组中观察到的高编码密度-这意味着大多数核苷酸读取将包含至少一部分蛋白质；(b)遗传密码的冗余——这减轻了核苷酸水平多态性的影响，这种多态性极大地混淆了核苷酸的组装，并且即使在潜在的组成核苷酸序列不相同的情况下，也可以重建蛋白质序列。肽组装器输出将用于开发一个框架，用于分析和比较基于其蛋白质和途径丰度的宏基因组样品。用于组装和分析的开源文档软件包将由该项目创建，并可供研究社区使用。这个项目将为宏基因组信息学的广泛领域的培训创造基础设施和工具，使从事生物科学研究的更广泛的社区以及数学和科学教育能够获得数据分析概念。此外，博士后和实习生将接受宏基因组学和计算生物学方面的培训。还将创建一个教育模块，通过研讨会向高中教师介绍基因组学和宏基因组学的生物信息学方法。参加讲习班的教师随后将在课堂上讲授课程。这种方法将促进年轻科学家发现和研究的兴奋。

This award to the J. Craig Venter Institute (JCVI) is to develop novel methods for analyzing metagenomic sequence data. Metagenomics pertains to the study of the genomic content of microbial communities using cultivation independent techniques. This paradigm has revolutionized the field of microbiology, especially given our inability to cultivate a majority of microbes that exist in many environments. Next-generation sequencing technologies are used routinely for generating large volumes of nucleotide sequence read data from metagenomic samples. The identification of full-length protein sequences from these data allows for a comprehensive and accurate analysis of the metabolic potential of the constituent microbes. This is often implemented by assembling the nucleotide sequence reads, and then using the set of generated contigs as substrate for protein identification. However, metagenomic assemblies are typically very fragmented, producing short contigs and also leaving a large fraction of reads unassembled, thereby limiting the utility of this approach. This project will develop a sequence assembly framework for reconstructing full-length protein sequences directly from short peptide fragments identified on nucleotide reads. This approach is motivated by two observations ? (a) the high coding density observed in prokaryotic genomes - which implies that most of the nucleotide reads will contain at least part of a protein; and (b) the redundancy in the genetic code - which alleviates the effect of nucleotide-level polymorphisms that greatly confound nucleotide assembly, and allows for the reconstruction of protein sequences even when the underlying constituent nucleotide sequences are not identical. The peptide assembler output will be used to develop a framework for analyzing and comparing metagenomic samples based on their protein and pathway abundances. Open-source documented software packages for assembly and analysis will be created by this project, and made available for use by the research community.This project will create infrastructure and tools for training in the broad area of metagenomic informatics to make data analysis concepts accessible to the wider community doing research in the biological sciences, and for mathematics and science education. Additionally, postdoctoral researchers and interns will be trained in metagenomics and computational biology. An educational module will also be created to introduce high school teachers, via workshops, to bioinformatic methods in genomics and metagenomics. The teachers participating in the workshops will subsequently teach the curriculum in their classrooms. This approach will promote the excitement of discovery and research in young scientists.