HyPhy: comprehensive, fast, and user-friendly software for evolutionary analysis

 描述（由申请人提供）：HyPhy（：www.hyphy.org）是一个可编写脚本的软件平台，旨在实现在进化背景下对DNA、RNA、密码子、氨基酸和其他类型的序列数据进行灵活而强大的分析。自2005年首次正式发布以来，HyPhy及其免费Web服务器Data Monkey（：www.datamonkey.org）已经变得稳定和成熟，被2，000多个同行评审的出版物引用，在多个书籍章节中描述，并成为许多受邀研讨会的主题。在第一个四年的供资周期中，我们侧重于改进和加强HyPhy系统的“后端”要素。在本提案中，虽然我们仍将在该级别上开展工作，但我们更侧重于“前端”， 与用户的互动是关键。目标1：重新设计HyPhy分析引擎及其脚本语言1.1重新设计HyPhy批处理语言（HBL）以提高其生产率、可靠性、可维护性、可移植性和可重用性;这将在保持向后兼容性的同时完成。1.2重新开发进化模型、标准分析和推理程序的标准库，使其具有自记录性、易于学习、易于扩展、对无意的误用具有鲁棒性，并符合数据交换格式和通信协议。目标2：大型复杂数据集的模型和算法。2.1通过加速硬件和软件中的基本操作，提高HyPhy性能，以便在一次分析中处理更大的数据集。2.2允许用户将联合收割机序列和其他定量数据组合在一个可能性框架中。2.3为快速演变的病原体和免疫库建立分析库。目标3：所有计算设备的基于Web浏览器的图形界面。目前，大多数HyPhy用户通过www.datamonkey.org与程序进行交互。考虑到对这种用户体验的需求，我们将：3.1实现一个完整的界面，用于数据探索，分析定义，作业执行，以及结果解释和可视化作为本地Web应用程序。这个界面将在计算机、桌子和智能手机上运行。3.2将HyPhy的计算核心开发为本地浏览器插件，使HyPhy成为一个完全在浏览器中分发、维护和运行的“应用程序”。3.3使用现代Web技术（node.js）重新实现datamonkey.org;提供一个可以从任何HyPhy实例访问的公共实例，以及一个允许实验室设置自己的基于集群或云的实例的分发版本。

 DESCRIPTION (provided by applicant): HyPhy (://www.hyphy.org) is a scriptable software platform designed to enable flexible and powerful analyses of DNA, RNA, codon, amino acid and other types of sequence data in an evolutionary context. Since their initial formal release in 2005, HyPhy and it's free web server Data Monkey (://www.datamonkey.org) have become stable and mature, cited in over 2,000 peer-reviewed publications, described in multiple book chapters, and been the subject of many invited workshops. In the first four-year funding cycle we focused on improvements and hardening of the "back end" elements of the HyPhy system. In this proposal, while we will still do work at that level, we focus more on the "front end" where interactions with users are key. Aim 1: Re-architect the HyPhy analytical engine and its scripting language 1.1 Redesign the HyPhy Batch Language (HBL) to enhance its productivity, reliability, maintain- ability, portability and reusability; this will be done while maintaining backward compatibility. 1.2 Redevelop the standard library of evolutionary models, standard analyses, and inference procedures to make them self-documenting, easy to learn, easy to extend, robust to inadvertent misapplication, and compliant with data exchange formats and communication protocols. Aim 2: Models and algorithms for large and complex datasets. 2.1 Improve HyPhy performance to handle much larger data sets in a single analysis by accelerating the fundamental operations in hardware and software. 2.2 Allow users to combine sequence and other quantitative data in a likelihood framework. 2.3 Develop a library of analyses for rapidly evolving pathogens and immune repertoires. Aim 3: Web browser based graphical interface for all computing devices. Presently, the majority of HyPhy users interact with the program through www.datamonkey.org. Keeping in mind the demand for such a user experience, we will: 3.1 Implement a complete interface for data exploration, analysis definition, job execution, and result interpretation and visualization as a local web application. This interface will run on computers, tables, and smartphones. 3.2 Develop the computational core of HyPhy as a native browser plug-in, making HyPhy an "app" distributed, maintained, and run entirely in a browser. 3.3 Re-implement datamonkey.org using modern web technologies (node.js); provide a public instance which can be accessed from any instance of HyPhy, and a distribution that can allow labs to set up their own cluster- or cloud-based instances.