Knowledge infrastructure for integrated data management and analysis supporting new approach methods in predictive toxicology and risk assessment

Barry Hardy,Tomaz Mohoric,Thomas Exner,Joh Dokler,Maja Brajnik,Daniel Bachler,Ody Mbegbu,Nora Kleisli,Lucian Farcal,Krzysztof Maciejczuk,Haris Rašidagić,Ghada Tagorti,Pascal Ankli,Daniel Burgwinkel,Divanshu Anand,Ugis Sarkans,Awais Athar
DOI: https://doi.org/10.1016/j.tiv.2024.105903
Abstract:The EU-ToxRisk project (2016-2021) was a large European project working towards shifting toxicological testing away from animal tests, towards a toxicological assessment based on comprehensive mechanistic understanding of cause-consequence relationships of chemical adverse effects. More than 40 partners from scientific institutions, industry and regulators coordinated their work towards this goal in a six-year long programme. The breadth and variety of data and knowledge generated, presented a challenging data management landscape. Here, we describe our approach to data management as developed under EU-ToxRisk. The main building blocks of the data infrastructure are: 1) An easy-to-use, extensible data and metadata format; 2) A flexible system with protocols for data capture and sharing from the entire consortium; 3) A methods database for describing and reviewing data generation and processing protocols; 4) Data archiving using a sustainable resource; 5) Data transformation from the archive to the system that provides granular access; 6) Application Programming Interface (API) for access to individual data points; 7) Data exploration and analysis modules, based on a «web notebook» approach to executable data processing documentation; and 8) Knowledge portal that ties together all of the above and provides a collaboration space for information exchange across the consortium. This knowledge infrastructure is being extended and refined for the support of follow-up projects (RISK-HUNT3R, ASPIS cluster, European Open Science Cloud (2021-2026)).
What problem does this paper attempt to address?