Metadata stewardship in nanosafety research: learning from the past, preparing for an "on-the-fly" FAIR future

Introduction: Significant progress has been made in terms of best practice in research data management for nanosafety. Some of the underlying approaches to date are, however, overly focussed on the needs of specific research projects or aligned to a single data repository, and this "silo" approach is hampering their general adoption by the broader research community and individual labs.Methods: State-of-the-art data/knowledge collection, curation management FAIrification, and sharing solutions applied in the nanosafety field are reviewed focusing on unique features, which should be generalised and integrated into a functional FAIRification ecosystem that addresses the needs of both data generators and data (re)users.Results: The development of data capture templates has focussed on standardised single-endpoint Test Guidelines, which does not reflect the complexity of real laboratory processes, where multiple assays are interlinked into an overall study, and where non-standardised assays are developed to address novel research questions and probe mechanistic processes to generate the basis for read-across from one nanomaterial to another. By focussing on the needs of data providers and data users, we identify how existing tools and approaches can be re-framed to enable "on-the-fly" (meta) data definition, data capture, curation and FAIRification, that are sufficiently flexible to address the complexity in nanosafety research, yet harmonised enough to facilitate integration of datasets from different sources generated for different research purposes. By mapping the available tools for nanomaterials safety research (including nanomaterials characterisation, nonstandard (mechanistic-focussed) methods, measurement principles and experimental setup, environmental fate and requirements from new research foci such as safe and sustainable by design), a strategy for integration and bridging between silos is presented. The NanoCommons KnowledgeBase has shown how data from different sources can be integrated into a one-stop shop for searching, browsing and accessing data (without copying), and thus how to break the boundaries between data silos.Discussion: The next steps are to generalise the approach by defining a process to build consensus (meta)data standards, develop solutions to make (meta)data more machine actionable (on the fly ontology development) and establish a distributed FAIR data ecosystem maintained by the community beyond specific projects. Since other multidisciplinary domains might also struggle with data silofication, the learnings presented here may be transferrable to facilitate data sharing within other communities and support harmonization of approaches across disciplines to prepare the ground for cross-domain interoperability.