de Sousa Ailton Lopes, Maués Dener, Lobato Amália, Franco Edian F, Pinheiro Kenny, Araújo Fabrício, Pantoja Yan, da Costa da Silva Artur Luiz, Morais Jefferson, Ramos Rommel T J
Institute of Biological Sciences, Federal University of Para, Belém, Brazil.
Institute of Exact and Natural Sciences, Federal University of Para, Belém, Brazil.
Front Genet. 2018 Dec 18;9:644. doi: 10.3389/fgene.2018.00644. eCollection 2018.
This study developed a computational tool with a graphical interface and a web-service that allows the identification of phage regions through homology search and gene clustering. It uses G+C content variation evaluation and tRNA prediction sites as evidence to reinforce the presence of prophages in indeterminate regions. Also, it performs the functional characterization of the prophages regions through data integration of biological databases. The performance of PhageWeb was compared to other available tools (PHASTER, Prophinder, and PhiSpy) using Sensitivity (Sn) and Positive Predictive Value (PPV) tests. As a reference for the tests, more than 80 manually annotated genomes were used. In the PhageWeb analysis, the Sn index was 86.1% and the PPV was approximately 87%, while the second best tool presented Sn and PPV values of 83.3 and 86.5%, respectively. These numbers allowed us to observe a greater precision in the regions identified by PhageWeb while compared to other prediction tools submitted to the same tests. Additionally, PhageWeb was much faster than the other computational alternatives, decreasing the processing time to approximately one-ninth of the time required by the second best software. PhageWeb is freely available at http://computationalbiology.ufpa.br/phageweb.
本研究开发了一种具有图形界面和网络服务的计算工具，该工具可通过同源性搜索和基因聚类来识别噬菌体区域。它使用G+C含量变异评估和tRNA预测位点作为证据，以加强在不确定区域中前噬菌体的存在。此外，它通过生物数据库的数据整合对前噬菌体区域进行功能表征。使用敏感性（Sn）和阳性预测值（PPV）测试，将PhageWeb的性能与其他可用工具（PHASTER、Prophinder和PhiSpy）进行了比较。作为测试的参考，使用了80多个经过人工注释的基因组。在PhageWeb分析中，Sn指数为86.1%，PPV约为87%，而第二好的工具的Sn和PPV值分别为83.3和86.5%。与接受相同测试的其他预测工具相比，这些数据使我们能够观察到PhageWeb识别的区域具有更高的精度。此外，PhageWeb比其他计算方法快得多，将处理时间减少到第二好的软件所需时间的约九分之一。PhageWeb可在http://computationalbiology.ufpa.br/phageweb免费获取。
