CRCNS US-German Data Sharing: DataGit - converging catalogues, warehouses, and deployment logistics into a federated 'data distribution'

当代神经科学在很大程度上是数据驱动的，但今天的数据管理技术和共享实践至少落后于软件生态系统同行十年。 分布式版本控制系统，如Git，促进了协作软件开发，而交钥匙发行版，如NeuroDebian，将研究人员从繁琐和不可靠的维护任务中解放出来。 同样，神经科学家也需要结合最新的技术发展，以更有效地访问、管理和反馈不断增长的科学数据。 通过一个简单的统一界面提供丰富的不相交数据集，可以超越单个研究的局限性，并彻底改变科学数据在所有科学领域的管理、分发和共享方式。 在美国国家科学基金会的支持下，雅罗斯拉夫·O.达特茅斯学院的Halchenko和沿着马格德堡大学（德国）的Michael Hanke博士将开发一套数据分发工具DataGit。 DataGit将采用数据跟踪和部署物流软件，统一访问许多现有的神经成像数据托管门户网站，如crcns.org、openfmri.org和humanconnectome.org。DataGit将使访问现有数据和共享新数据或派生数据变得容易，并完全支持分布式版本控制、数据完整性保护和对原始数据托管的身份验证访问。使数据管理像源代码管理一样简单和通用，将进一步推动开放和完全可复制的科学。对联合数据集的统一访问将促进该领域内外神经科学数据的可见性和可访问性，远远超出任何个人数据共享工作的范围。 拟议的发展所带来的好处将直接转化为教育工作者在课堂上的目标。 通过与软件发行版的整合，统一获取在线培训材料的软件元素和数据集，将使教育工作者不仅能够从教科书中教学，而且还能够通过动手复制最先进的原始出版物进行教学。 此外，让任何研究人员都能够轻松部署复杂的异构分析管道，这将有助于将人类连接组项目等旗舰项目的成果转化为临床应用的可用工具。 因此，更多的研究人员将能够通过提高我们对人类大脑的理解来应对更大的挑战，从而造福社会。

Contemporary neuroscience is heavily data-driven, but today's data management technologies and sharing practices fall at least a decade behind software ecosystem counterparts. Distributed version control systems, such as Git, facilitate collaborative software development, and turnkey distributions, like NeuroDebian, free researchers from tedious and unreliable maintenance tasks. Likewise, neuroscientists will need to incorporate recent technological developments to access, manage, and contribute back to the ever growing array of scientific data more efficiently. Making a rich collection of disjoint datasets available through a simple unified interface can transcend limitations of individual studies and revolutionize how scientific data are managed, distributed, and shared across all fields of science. With support from the National Science Foundation, Dr. Yaroslav O. Halchenko of Dartmouth College, along with Dr. Michael Hanke of the University of Magdeburg (Germany), will develop DataGit, a suite of data distribution tools. DataGit will employ software for data tracking and deployment logistics to unify access to many existing neuroimaging data hosting portals, such as crcns.org, openfmri.org and humanconnectome.org. DataGit will make it easy to access existing data and to share new or derived data with full support for distributed version control, data integrity protection and authenticated access to original data hosting.Making data management as easy and as versatile as source code management will further the efforts toward open and fully reproducible science. Uniform access to federated collections of data will promote the visibility and accessibility of neuroscientific data inside and outside the field, far beyond the scope of any individual data-sharing effort. The benefits from the proposed developments will translate directly to educators' aims in the classroom. Through integration with software distributions, uniform access to software elements and datasets for online training materials will enable educators to teach not only from textbooks but also through hands-on replication of state-of-the-art original publications. In addition, giving any researcher the ability to easily deploy complex heterogeneous analysis pipelines will be instrumental in translating the achievements of flagship efforts, such as the Human Connectome Project, into accessible tools for clinical applications. Consequently, even more researchers will be able to tackle even larger challenges to benefit society by improving our understanding of the human brain.A companion project is being funded by the German Ministry of Education and Research (BMBF).