Modern scientific data management requires comprehensive support for the FAIR (Findable, Accessible, Interoperable, Reusable) principles. Here we demonstrate a general approach, or design pattern, to supporting FAIR principles for diverse neuroscience data. The design pattern ensures that the key scientific and technical activities and agents of the data generation process are expressed in a validatable provenance-based data model. Thus, the data models capture contextual information necessary to interpret the scientific meaning of the data, infer the resulting data types, evaluate trust and quality, ensure attribution of all contributors, and support data reuse, integration, interoperability and longevity. We describe the principles of the design pattern and demonstrate its generality using multiple data types including neuron morphologies, electrophysiological data, brain atlases and computational models. These data models are currently used to support search, discovery, provenance tracking, publishing and data-driven modeling workflows in the Blue Brain Project and the EU Human Brain Project. A public repository (http://github.com/INCF/neuroshapes) and an INCF Special Interest Group (https://www.incf.org/activities/standards-and-best-practices/incf-special-interest-groups/incf-sig-on-neuroshapes-open) have been established to disseminate and foster collaboration on these data models. This design pattern may be valuable for supporting FAIR data principles in other domains.
The main goal is to promote: * the use of standard semantic markups and linked data principles as ways to structure metadata and related data: the W3C RDF format is leveraged, specifically its developer-friendly JSON-LD serialization. The adoption of linked data principles and JSON-LD will ease federated access and discoverability of distributed neuroscience (meta)data over the web. * the use of the W3C SHACL (Shapes Constraint Language) recommendation as a rich metadata schema language which is formal and expressive; interoperable; machine-readable; and domain-agnostic. With SHACL, (meta)data quality can be enforced based on schemas and vocabularies (easily discoverable and searchable) rather than being fully encoded in procedural codes. SHACL also provides key interoperability capabilities to ensure the evolution of standard data models and data longevity. It allows to incrementally build standard data models in terms of semantics and sophistication. * the reuse of existing schemas and semantic markups (like schema.org ) and existing ontologies and controlled vocabularies (including NIFSTD - NIF Standard Ontologies) * the use of the W3C PROV-O recommendation as a format to record (meta)data provenance: a SHACL version of the W3C PROV-O is created.
+ Modern scientific data management requires comprehensive support for the FAIR (Findable, Accessible, + Interoperable, Reusable) principles. Here we demonstrate a general approach, or design pattern, to + supporting FAIR principles for diverse neuroscience data. The design pattern ensures that the key + scientific and technical activities and agents of the data generation process are expressed in a + validatable provenance-based data model. Thus, the data models capture contextual information necessary + to interpret the scientific meaning of the data, infer the resulting data types, evaluate trust and + quality, ensure attribution of all contributors, and support data reuse, integration, interoperability + and longevity. We describe the principles of the design pattern and demonstrate its generality using + multiple data types including neuron morphologies, electrophysiological data, brain atlases and + computational models. These data models are currently used to support search, discovery, provenance + tracking, publishing and data-driven modeling workflows in the Blue Brain Project and the EU Human + Brain Project. A public repository (http://github.com/INCF/neuroshapes) and an INCF Special Interest + Group + (https://www.incf.org/activities/standards-and-best-practices/incf-special-interest-groups/incf-sig-on-neuroshapes-open) + have been established to disseminate and foster collaboration on these data models. This design pattern + may be valuable for supporting FAIR data principles in other domains. +
++ The main goal is to promote: + * the use of standard semantic markups and linked data principles as ways to structure metadata and related data: the W3C RDF format is leveraged, specifically its developer-friendly JSON-LD serialization. The adoption of linked data principles and JSON-LD will ease federated access and discoverability of distributed neuroscience (meta)data over the web. + * the use of the W3C SHACL (Shapes Constraint Language) recommendation as a rich metadata schema language which is formal and expressive; interoperable; machine-readable; and domain-agnostic. With SHACL, (meta)data quality can be enforced based on schemas and vocabularies (easily discoverable and searchable) rather than being fully encoded in procedural codes. SHACL also provides key interoperability capabilities to ensure the evolution of standard data models and data longevity. It allows to incrementally build standard data models in terms of semantics and sophistication. + * the reuse of existing schemas and semantic markups (like schema.org ) and existing ontologies and controlled vocabularies (including NIFSTD - NIF Standard Ontologies) + * the use of the W3C PROV-O recommendation as a format to record (meta)data provenance: a SHACL version of the W3C PROV-O is created. +
+