WorldFAIR (D4.2) FAIRification of nanoinformatics tools and models recommendations

Nanomaterials, with properties of both chemicals and particles, offer exciting opportunities in a range of industrial and consumer applications, from sensing and diagnostics to precision medicine and agriculture. Paradoxically, the properties that make then advantageous for applications, including their small size and large surface area, are also the source of concerns regarding potential negative impacts arising from their uptake by, and interactions with, humans and the environment. Given the enormous diversity of nanomaterials compositions, it is not possible to individually test them using the current time, cost and animal-intensive regulatory testing approaches, driving an urgent need for alternative in silico approaches to predict nanomaterials safety (nanoinformatics).  

The nanomaterials safety community have been actively developing a range of modelling approaches, spanning from physics-based models to data-driven approaches including machine learning models. As these models utilise and generate extensive datasets, there is a requirement for good practice in data documentation to support model development. Additionally, the models and associated software need to be FAIR (findable, accessible, interoperable and re-usable). While there is much in common with the FAIR needs for software in chemoinformatics, there are some unique aspects to nanomaterials software (nanoinformatics) that require domain-specific tailoring.

This WorldFAIR Deliverable report, which is targeted towards nanoinformatics model developers, presents a set of recommendations and prototypes for FAIRification of nanoinformatics tools and models. The deliverable is a stand-alone document focused on FAIRification of nanoinformatics tools and software primarily, addressing also FAIRIfication of the underpinning (and resulting) datasets. Organisation of the datasets into ready-for-modelling formats, for example via NanoPharos, and use of KNIME nodes to integrate the datasets directly into the modelling software, and the resulting predictions and validation statistics back into the database for further re-use are also emphasised.   

This report provides an analysis of the direction FAIRification of nanoinformatics software could/should take. The report provides examples of the approaches and best practice that have emerged from Horizon 2020-funded (H2020) nanosafety-specific projects including NanoCommons, NanoSolveIT, RiskGONE, CompSafeNano to support FAIR software. Approaches and best practice examples include the documentation of models and software via existing and emerging metadata standards, establishment of a registry of nanoinformatics models, deployment of predictive models as web applications or application programming interfaces and a demonstration of model interoperability and enhanced re-usability via containerisation and deployment via a cloud platform. Recommendations for next steps are provided to drive progress.

Visit WorldFAIR online at http://worldfair-project.eu.

WorldFAIR is funded by the EC HORIZON-WIDERA-2021-ERA-01-41 Coordination and Support Action under Grant Agreement No. 101058393.