SciDAP: a next generation universal platform for collaborative data analysis

最近，基于下一代测序(NGS)的基因表达分析方法的激增-- Sion、染色质结构和蛋白质-DNA相互作用为获得新的 对基础生物学、健康和疾病的洞察。然而，对结果数据的分析需要计算 许多传统生物学家不具备的专业知识。因此，在处理基因组数据时，大多数 即使是简单的任务，生物学家也需要生物信息学家的帮助。这使这些令人兴奋的方法超越了 大多数生命科学家的触角。 来自俄亥俄州辛辛那提的初创公司DATIRIUM，LLC的这份提案描述了一项创建 SciDAP(科学数据分析平台)，一个新的多组学用户友好的数据分析平台，允许 生物学家分析数据本身，并使生物学家和生物信息学家之间的合作成为可能。 Datirium是由该应用程序的PI Artem Barski、PhD和Andrey Kartashov创建的，最初是为了帮助用户 通过BioWardrobe的安装和支持。BioWardrobe，一个用户友好的开源集成基因组学 分析平台，由辛辛那提儿童医院医学中心巴斯基实验室开发 (CCHMC)在2015年。它已被40多个CCHMC实验室使用，处理了8000多个实验。 目前已在40多种出版物中得到应用。此外，Datirium已经安装并继续 在几个外部研究中心维护BioWardrobe服务器。 对于Datirium来说，BioWardrobe作为最低限度的可行产品(MVP)，允许确认需要 以及此类软件存在的利基市场，但也突显了几个设计上的局限性。其中的关键 它们是增加新管道或修改现有管道的困难：由于管道之间的紧密耦合 和用户界面，这需要在所有级别的软件上进行更改。不幸的是，所有人都存在同样的限制 用户友好的生物信息学工具。假设有150多种基于NGS的方法和许多方法 在处理数据方面，这解释了为什么还没有一个通用的、用户友好的数据分析平台。 我们假设我们可以通过包含以下内容来创建一个既通用又用户友好的数据分析平台- 将接口指令插入计算流水线本身。Platform将使用这些说明创建 为用户提供图形界面。具体地说，我们正在使用使用Common Work开发的集装箱化管道- 流语言(CWL)。CWL允许描述制作这些的工具、管道和计算环境 管道既可携带又可重复使用。在CWL之上，Datirium开发了一个CWL扩展系统 允许描述CWL工作流中的输入和输出可视化。重要的是，我们的平台 将通过以下方式增加计算分析的严谨性：(I)通过生物素使分析可重现和审计- 由于CWL管道的便携性而形成的形式学家，并将分析的每一步记录为研究对象； 通过提供生物信息学家实现实验者和计算生物学家之间的协作 通过图形用户界面方便地指导分析流程和生物学家；(Iii)包括开箱即用管道- 具有优化的参数和可操作的QC指标的行，以标记可能的问题。 在本提案的第一个目标中，我们将创建该平台的原型。在第二个目标中，我们将进行 与生物信息学家和实验者一起进行可用性测试，以测试我们的平台是否能够 以用户友好的方式进行多种类型的分析。具体地说，与Salomonis博士在 CCHMC，我们将测试将两个现有的分析例程集成到SciDAP中有多容易：bs-Seq DNA-甲基- 和scRNA-Seq；然后我们将与生物学家合作，确保得到的界面既用户友好 并实现了与生物信息学家的合作。 该项目的成功完成将为研究界提供尖端、灵活和 生物学家友好的数据分析平台。

The recent proliferation of next-generation sequencing (NGS) - based methods for the analysis of gene expres- sion, chromatin structure and protein-DNA interactions has created tremendous opportunities for gaining novel insights into basic biology, health, and disease. However, analysis of the resulting data requires computational expertise that many traditional biologists do not possess. Hence, when dealing with genomics data, majority of biologists require the help of bioinformaticians even for simple tasks. This places these exciting methods beyond the reach of the majority of life scientists. This proposal from DATIRIUM, LLC, a start-up company from Cincinnati, OH describes a plan to create SciDAP (Scientific Data Analysis Platform), a novel multi-omics user-friendly data analysis platform to allow biologists to analyze the data themselves and to enable collaboration between biologists and bioinformaticians. Datirium was founded by this application’s PIs, Artem Barski, PhD and Andrey Kartashov, initially to assist users with installation and support of BioWardrobe. BioWardrobe, a user-friendly open-source integrative genomics analysis platform, was developed by the Barski laboratory at Cincinnati Children’s Hospital Medical Center (CCHMC) in 2015. It has been used by more than 40 CCHMC laboratories to process more than 8000 experi- ments and has been applied in more than 40 publications. In addition, Datirium has installed and continues to maintain BioWardrobe servers at several external research centers. For Datirium, BioWardrobe served as a Minimum Viable Product (MVP) that allowed to confirm the need and existence of market niche for such software, but also highlighted several design limitations. The key among them was the difficulty in adding new or modifying existing pipelines: due to the tight coupling between pipeline and user interface this required changes at all levels of software. Unfortunately, the same limitation exists for all user-friendly bioinformatics tools. Given that there are more than 150 NGS-based methods and many ways to process the data, this explains why a universal and user-friendly data analysis platform does not yet exist. We hypothesize that we can create a data analysis platform that is both universal and user-friendly by includ- ing interface instructions into computational pipelines itself. Platform will use these instructions to create a graphical interface for users. Specifically, we are using containerized pipelines developed using Common Work- flow Language (CWL). CWL allows to describe tools, pipelines and computational environment making these pipelines both portable and reproducible. On top of CWL, Datirium developed a system of CWL extensions that allows to describe the inputs and outputs visualizations within the CWL workflows. Importantly, our platform will increase the rigor of computational analysis by (i) making the analysis reproducible and auditable by bioin- formaticians due to CWL pipeline portability and recording each step of the analysis as Research Objects; (ii) enabling collaboration between experimentalists and computational biologists by providing bioinformaticians with a way to direct analysis flow and biologists with the convenience of GUI; (iii) Including out of the box pipe- lines with optimized parameters and actionable QC metrics that flag possible issues. In the first aim of this proposal we will create a prototype of the platform. In the second aim, we will conduct usability testing with bioinformaticians and experimentalists to test whether our platform can accommodate diverse types of the analysis in a user-friendly fashion. Specifically, in collaboration with Dr. Salomonis at CCHMC, we will test how easy it is to integrate two existing analysis routines into SciDAP: BS-Seq DNA-methyl- ation and scRNA-Seq; then we will work with biologists to ensure that the resulting interface is both user-friendly and enables collaboration with bioinformaticians. Successful completion of this project will provide the research community with a cutting edge, flexible and biologist-friendly data analysis platform.