The EBRAINS Knowledge Graph (KG) integrates registered data, models, and software into a graph database, and outlines their relationship to each other, i.e. allowing you to connect a dataset to software tools for analysis and visualisation.
The knowledge graph technology is used in many organisations and for many purposes. It organises and links information from various sources in a web-like structure for the purpose of defining and exploring relationships and generating knowledge. Knowledge graphs have properties that make them more suitable than conventional fixed-schema databases in many application areas:
they allow existing data models to be extended at any point in time with new properties and connections
they can be used to generate a graphical representation of the relationships between any of its data points, allowing dynamic and flexible navigation of the content
complex relationships between data points can be represented more easily and thanks to semantic annotations, new connections can be inferred
Adopting the knowledge graph technology to a huge and diverse field of research such as neuroscience does not come for free. In many other application areas, pre-processed information is fed into the graph structures via automated ingestion pipelines. For the field of neuroscience, such pipelines are in most cases not easy to establish since the pre-processed information is either not available or not possible to interpret for others than those who generated the information.
The EBRAINS KG addresses this challenge through an “expert-driven” approach. Neuroscience-specific metadata standards and conventions are introduced while supporting extensions and adaption for future needs. The design employed makes conflicting information transparent and therefore enforces scientific discussions. This approach (which we call “expert-driven”) adds several challenges which are addressed by various components built on top of the core services of the Knowledge Graph.
Ingested data are controlled by manual, semi-automated and automated quality assurance processes allowing domain-specific experts (in this case the neuroscientists of the EBRAINS Curation Services to actively validate and ensure conventions and consistency of the data.