The produced data artifacts in the Human Brain Project are quite heterogeneous and the requirements to appropriately reflect complex scientific protocols while trying to unify the annotations across different domains is challenging.
To make meta-data (and therefore the underlying data) findable, accessible, interoperable and reproducible (FAIR), the HBP relies on a professional curation support which ensures a high quality of the meta-data and the use of a "common language" (e.g. ontologies, terminologies, etc.).
The HBP Knowledge Graph supports this endeavour by providing the toolset for the curation workflows by following a "open world" approach and making heavy use of standards of the semantic web (JSON-LD):
Being permissive in terms of what meta-data can be registered is an absolute necessity to allow the different scientific protocols, processes and activities to be appropriately represented in the Knowledge Graph. The open world approach means, that existing data models can be extended at any moment in time with new properties and connections opposed to conventional, fixed-schema data structures.
At the same time it is important to follow conventions of how different reoccuring elements should be described (e.g. which fields shall be provided, which ontologies shall be used, etc.).
The system allows the experts of the professional curation support to define the standards, ontologies and terminologies they agree on. Once defined, the provided tools support its users to comply with the conventions via suggestion systems.
In a highly dynamic field like neuroscience, even the most carefully prepared and established annotation conventions combined with a professional curation pipeline will not prevent data structures to evolve over time: New fields are added, new connections built, values and ontologies introduced and/or deprecated, etc.
The HBP Knowledge Graph is aware of this fact and provides tools and mechanisms which incorporate the organizations conventions while supporting the extension, adaptation and migration of meta-data and its structures.
Although a lot of the technical work is done "behind the scenes", the KG Editor is the most visible component of the meta-data management system provided of the solution.
Find more information depending on:
Or contact us by e-mail: email@example.com