Computational Annotation of Orphan Metabolic Activities

描述：大量的测序和基因组数据有可能彻底改变我们对微生物生物学的理解，并改变医学研究。细菌新陈代谢的准确重建通常为理解测序物种的生物学、它们的生长和环境特性以及与元基因组群落中其他物种的可能相互作用提供了最直接的途径。重建的代谢网络也可以指导新的抗菌治疗药物的开发。在这个应用中，我们建议使用Globus框架来建立一个集成的、准确的和完全概率的系统来注释细菌代谢；我们将把关键的数据模式集成到框架中，根据我们的初步结果，这些数据模式将显著提高方法的覆盖率和准确性。具体地说，我们将。)利用酶活性位点分析，将蛋白质结构信息整合到Globus中；b)通过对代谢组学和蛋白质注释的联合采样，扩展Globus框架以容纳代谢组学数据；将表型信息整合到注释框架中，包括广泛使用的用于高通量细菌表型鉴定的Biolog平台测量的多种营养源上的生长情况。我们将实施开发的方法，并通过一个透明的基于网络的门户网站向公众提供。这将使其他研究人员有可能使用Globus方法来注释任何感兴趣的已测序细菌物种。门户将是透明的，使用户能够识别预测的来源我们还将从我们的实验合作者那里获得相关的表型信息，并使用Globus为所有对人类致病的主要细菌物种(~50个细菌)生成准确的概率注释。

DESCRIPTION: An avalanche of sequencing and genomic data has a potential to revolutionize our understanding of microbial biology and transform medical research. Accurate reconstructions of bacterial metabolism often provide the most direct route toward understanding the biology of sequenced species, their growth and environmental properties, and possible interactions with other species in metagenomic communities. Reconstructed metabolic networks can also guide the development of new antibacterial therapeutics. In this application we propose to use the GLOBUS framework to build an integrated, accurate, and fully probabilistic system for the annotation of bacterial metabolism; we will integrate into the framework key data modalities that, according to our preliminary results, will significantly improve the method's coverage and accuracy. Specifically, we will a.) integrate protein structural information into GLOBUS, using analyses of enzyme active sites; b.) extend the GLOBUS framework to accommodate metabolomics data by joint sampling of metabolomics and protein annotations; and c.) integrate into the annotation framework phenotypic information, including growth on multiple nutrient sources measured by the widely used BiOLOG platform for high-throughput bacterial phenotyping. We will implement and make the developed methodology publically available through a transparent web-based portal. This will make it possible for other researchers to use the GLOBUS methodology to annotate any sequenced bacterial species of interest. The portal will be transparent, enabling users to identify the sources of the predictions We will also obtain relevant phenotypic information from our experimental collaborators and use GLOBUS to generate accurate probabilistic annotations for all major bacterial species (~50 bacteria) that are pathogenic to humans.