Supplement: Enhancing Community Contributions to Bioconductor With Build System Containerization and a GPU for Testing

项目摘要/摘要 《BioConductor：一种开源、开放开发的基因组计算资源》是 此补充基金申请的家长批准书。BioConductor是一个开源的、开放的 开发用于分析和理解基因组的软件和数据生态系统- 规模实验。该系统主要植根于R语言，但具有广泛的 互操作性功能，并整合了在许多其他 语言。在整个科学和工业场所都需要生物导体资源 从BioConductor的云分发系统流出的大约1TB 每天的软件和数据。该系统为基因组表示和 众多模型的变体、基因组序列和基因模型的表示 生物体，用于许多阵列和测序平台的分析软件，以及注释和 实验档案，每天都会检查其一致性和便携性。此请求 补充资金用于满足建造和分配的工程要求 系统，它目前的结构已经有十多年了，并增加了对图形的支持 构建和测试系统中的处理单元。我们的计划包括集装箱运输和 扩大建立和检查系统的规模，改进记录方法 生态系统建设过程中的资源使用和不良事件。大体上，拟议的 这项工作将使BioConductor项目能够更全面地接受“基础设施即代码” 简化重新部署和重新调整以响应吞吐量要求的纪律，以及 改进系统问题的自动发现和报告。因为这一要求 基因组数据科学界正在不断地在复杂性和 生成的数据量和正在使用的硬件平台的数量也是如此 随着规模的不断扩大，拟议的工作不可能在最初的供资计划下完成。

PROJECT SUMMARY/ABSTRACT "Bioconductor: An Open-Source, Open-Development Computing Resource for Genomics" is the parent grant of this application for supplemental funds. Bioconductor is an open source, open development ecosystem of software and data for the analysis and comprehension of genome- scale experiments. The system is primarily rooted in the R language but has extensive interoperability capabilities and incorporates tools developed in numerous other languages. Bioconductor resources are requested at scientific and industrial sites throughout the world, with an outflow from Bioonductor's cloud distribution systems of approximately 1TB of software and data per day. The system provides infrastructure for genome representation and representation of variants, genomic sequences and gene models for numerous model organisms, analytical software for many array and sequencing platforms, and annotation and experiment archives, all checked for consistency and portability on a daily basis. This request for supplemental funds addresses engineering requirements of the build and distribution system, which has had its present structure for over ten years, and adds support for graphical processing units in the build and test system. Our plan includes work on containerization and upscaling of the build and check system, and enhancement of the methods of recording resource use and adverse events in ecosystem build processes. In essence, the proposed work will allow the Bioconductor project to more fully embrace an "Infrastructure as Code" discipline to simplify redeployment and rescaling in response to throughput requirements, and to improve automatic discovery and reporting of system problems. Because the requirements of the genomic data science community are continually diversifying in terms of complexity and volume of data being generated, and the number of hardware platforms in use is likewise growing, the work proposed cannot be accomplished under the initial funding plan.