Recommendations for repositories and scientific gateways from a neuroscience perspective | Scientific Data – Nature.com

Data repositories and scientific gateways have the potential to contribute strongly with technical reproducibility and consistent data quality. Unique identifiers make data easy to find and cite. Structured method reporting and automated metadata verification make data more reliable and reusable.

The use of (machine readable) persistent identifiers (PID) is a core requisite for making research data accessible and fulfilling the FAIR principles. Services should assign PIDs to data descriptions, data and complementary materials (e.g., digital object identifiers (DOI)), software (DOI, Software Heritage ID (SWHID)14), authors (open researcher and contributor IDs (ORCID)) and associated research resources (RRIDs15). We also recommend that service providers register for an RRID that identifies their infrastructure.

Metadata is critically important to FAIR6; it is the backbone of any dataset, and ongoing quality control of metadata is as important as the data. It is vital in ensuring that data can be correctly understood and effectively used and reused.

We recommend services to document and communicate their curation processes for data and metadata. Where possible, higher level curation which links to annotation and other published information material is preferable.

We recommend that methods are reported in a structured, community relevant format, (examples: Structured, Transparent, Accessible Reporting (STAR) Methods, MDAR (Materials Design Analysis Reporting)) and that metadata entry is made easy and automatically or semi-automatically verified. Ideally, methods are also published and citable (using platforms such as protocols.io).

We recommend that key software, such as analysis code, is versioned and documented, and that the versioning history is communicated. Provenance for data, derived data and software should be documented and extractable. We recommend that versioning of both content and authorship is transparently communicated and available for datasets, code, and analysis software.

We recommend services to interact with their community to identify and accommodate various data search behaviours, and to deliver search summaries that make it possible for researchers to judge relevance, accessibility, and reusability of a data collection from the summary.

Here is the original post:
Recommendations for repositories and scientific gateways from a neuroscience perspective | Scientific Data - Nature.com

Related Posts