Gast ::
| Hauptseite > Publikationsdatenbank > Workflows for metadata management in neuroscience |
| Hauptseite > Publikationsdatenbank > Workflows for metadata management in neuroscience |
| Poster (After Call) | FZJ-2022-05426 |
; ; ; ;
2022
Abstract: Tracking and documenting how data is generated, processed, and analyzed is crucial for scientists to replicate and reuse experimental results. However, the sheer quantity of the resulting metadata generated by today’s complex investigations can be difficult for researchers to record, manage, and use effectively. The Helmholtz Metadata Collaboration (HMC; https://helmholtz-metadaten.de) supports scientists of the Helmholtz Association in creating and handling metadata, to help make research results more sustainable. Within HMC, interactions with scientific disciplines are organized into six hubs. The Hub Information focuses on metadata topics in materials science, physics, computing, plant science, and neuroscience. We develop software and semantic architectures to help manage and use metadata, assess how FAIR (Findable, Accessible, Interoperable, and Reusable; [1]) research outputs are, and conduct education events for scientists. While we often work on specific use cases, our goal is always to produce solutions that can be generalized to other disciplines.During a typical research workflow, it is crucial that scientists are able to effectively (1) store and link data and metadata, then use this metadata to (2) inform analyses. To manage data and metadata in neuroscience, we use several existing and freely available software tools, and are developing additional tools to fill in the gaps. We link these tools together using a workflow developed with the open-source programming language Python. Here, we summarize our current approaches and future directions.To (1) store and link data and metadata, we currently prefer odML documents for metadata and NIX files for combining data and metadata. odML files store metadata in an xml-based format containing key-value pairs in a hierarchical structure – they support links between metadata items within a document, and are designed for storing metadata without data [2,3]. NIX files store raw and preprocessed data and metadata together in a single file, using a generic data structure and the odML model for metadata – they support links between data and metadata objects within a file [4]. Because a single NIX file contains all the information about an experiment, it is easy to store and to share individual data records with collaborators. The NIX data model standardizes access to data and metadata from different sources. For example, NIX files can map the neuroscience-specific Neo data representation [5], ensuring that data and metadata annotations are consistently accessible.We are finding, however, that many neuroscientists need to accurately describe experimental setups where relationships between elements are more nuanced than odML files can describe. To address this problem, we are developing specific ontologies and terminologies for aspects of experimental setups in brain electrophysiology. Ontologies are an alternative approach to represent and link metadata, and include precise descriptions of defined elements and their relationships. Consistent descriptions will help standardize metadata from different experiments, facilitating data reuse.After data and metadata have been stored, researchers face challenges in using metadata to (2) inform analyses because they lack tools to efficiently navigate the enormous amounts of metadata generated by research projects. To solve this problem, we are developing software to combine metadata from multiple experiments into a database which users can interactively explore by selecting different criteria. The software condenses metadata to show only information chosen by the user, in both a data table and a graphical overview. This gives users an overview of their experiments to inform further analyses.Making metadata descriptive and accessible ensures that others can reproduce and build upon research discoveries. We envision that our tools will help facilitate open and reproducible science by making metadata easier to collect and use.References:[1] Wilkinson, M.D., et al., 2016. The FAIR Guiding Principles for scientific data management and stewardship. Scientific Data 3, 160018. https://doi.org/10.1038/sdata.2016.18[2] Grewe, J., Wachtler, T., Benda, J., 2011. A bottom-up approach to data annotation in neurophysiology. Frontiers in Neuroinformatics 5, 16. https://doi.org/10.3389/fninf.2011.00016[3] Sprenger, J., Zehl, L., Pick, J., Sonntag, M., Grewe, J., Wachtler, T., Grün, S., Denker, M., 2019. odMLtables: A user-friendly approach for managing metadata of neurophysiological experiments. Front. Neuroinform. 13, 62. https://doi.org/10.3389/fninf.2019.00062[4] Stoewer A, Kellner CJ, Benda J, Wachtler T and Grewe J (2014). File format and library for neuroscience data and metadata. Front. Neuroinform. Conference Abstract: Neuroinformatics 2014. doi:10.3389/conf.fninf.2014.18.00027[5] Garcia S, Guarino D, Jaillet F, Jennings T, Pröpper R, Rautenberg PL, Rodgers CC, Sobolev A, Wachtler T, Yger P and Davison AP (2014). Neo: an object model for handling electrophysiology data in multiple formats. Frontiers in Neuroinformatics, 8, 10. http://doi.org/10.3389/fninf.2014.00010
|
The record appears in these collections: |