Metadata collection framework for consistent storage, analysis and collaboration

Abstract

Event Abstract Back to Event Metadata collection framework for consistent storage, analysis and collaboration Michael Sonntag1*, Adrian Stoewer1*, Andrey Sobolev1*, Cristina Precup1 and Thomas Wachtler1 1 Ludwig-Maximilians-Universität München, German Neuroinformatics Node, Germany Recent progress in neuroscience leads to increasingly complex protocols, experimental approaches, and variety in experimental metadata. Availability of tools for reliable metadata consolidation, as well as for effortless data and metadata access becomes crucial for efficient and reproducible research. In this work we present a framework targeted to improve metadata collection, storage, access and exchange, as important ingredients of experimental electrophysiology. The framework сomprises a set of tools [1] for consistent metadata management in a single database. Metadata are always kept aligned to the same object model, relevant for a particular domain of neuroscience. To account for the huge diversity of experimental settings, we use modern resource description framework techniques (RDF, [2]), which provide the required flexibility in data annotation while enabling consistent organization and machine-readability. A common data scheme is provided by a core ontology with generally used terms that can be extended and customized to fit the specific requirements of an individual lab or project. A flexible plugin system enables including tools to extract metadata from proprietary file formats for automated metadata collection. Data storage is file based, supporting distributed storage and integration from multiple users and versioning using popular tools like git [3]. In addition, the usage of RDF enables integration of standards for provenance tracking [4] into the metadata collection workflow. A graphical interface provides key functions to create, manage, search and query metadata and annotations, but one can also directly access the stored metadata in files. Metadata is saved using standard RDF formats accessible with open source RDF libraries from Python, C/C++, Matlab and other languages. Moreover, the framework provides Java-based application access (API). These options enable integrating metadata and data management seamlessly within the data analysis workflow, fostering scientific progress through neuroinformatics. Acknowledgements Supported by the German Federal Ministry of Education and Research (Grant 01GQ1302).