Shortening the development cycle time of Trans Proteomic Pipeline tools with high performance computing

项目摘要 基于质谱学的蛋白质组学是目前应用最广泛的蛋白质组学分析技术 复杂的蛋白质混合物。它有能力检测和量化数千种蛋白质的丰度和 它们的变体，翻译后修饰，蛋白质：蛋白质和蛋白质：每个实验的药物相互作用。这个 TPP是一套免费的、开源的质量分析软件工具 波谱(MS)蛋白质组学数据已经被社区广泛使用了超过15年。我们正在进行的 TPP维护和发展奖有四个目标，我们1)加强TPP以 处理新兴工作流中的数据集，2)为大型数据库添加统计分析和可视化功能 数据集，3)通过功能和可用性增强响应社区需求，以及4)扩展 我们的用户基础经过改进的培训和在线学习。在实现这些目标方面取得了很大进展 项目期。然而，我们目前关注的主要工具之一的当前开发已经 由于2012年购买的计算服务器老化和开发不足， 计算资源。我们建议通过购买和部署一对高性能的 内存、高核数、快速磁盘计算服务器，大大加快了我们的开发周期 可以更快地完成算法开发，并且可以进一步快速地完成算法 经过优化，通常以更适中的硬件要求运行。这些机器的部署将 通过提供有关软件变更的快速反馈，进一步简化我们的软件开发 发展及其对搜索速度的影响，使我们能够更快地向用户提供工具， 加快许多其他使用TPP进行数据处理的研究项目的进度。

Project Summary Mass spectrometry (MS) based proteomics is currently the most widely used technology for the analysis of complex protein mixtures. It has the ability to detect and quantify the abundance of thousands of proteins and their variants, post-translational modifications, protein:protein and protein:drug interactions per experiment. The Trans-Proteomic Pipeline (TPP) is a free and open-source suite of software tools for the analysis of mass spectrometry (MS) proteomics data that has been widely used by the community for over 15 years. Our ongoing award for the maintenance and development of the TPP has four aims in which we 1) enhance the TPP to process datasets from emerging workflows, 2) add statistical analysis and visualization capabilities for large datasets, 3) respond to the needs of the community with functional and usability enhancements, and 4) expand our user base with improved training and on-line learning. Great progress has been made on these aims during the project period. However, current development for one of the main tools that we are currently focusing on has been severely hindered by aging computational servers purchased in 2012 and inadequate development of computing resources. We propose to address this problem via the purchase and deployment of a pair of high- memory, high-core count, fast-disk compute servers, substantially accelerating our development cycle so that the algorithm development can be completed more rapidly, and the completed algorithm can be further rapidly optimized to run with more modest hardware requirements routinely. Deployment of these machines will streamline our software development by providing rapid feedback on software changes and further developments and their impact on search speed and enable us to provide tools to our users more rapidly, accelerating the progress of many other research programs that use the TPP for data processing.