A productive meeting in Tartu, Estonia, brought together modelers from across PARC to share progresses, exchange ideas, and build new collaborations.

The PARC Modelers' Meeting, held from June 9 to 11 2025, was hosted by the Chair of Molecular Technology of the Institute of Chemistry of the University of Tartu (Estonia). It gathered around 90 participants on site and online, including modelers and data managers from several PARC work packages.

The main goals of the meeting were to review progress, present future plans, and identify broader synergies in the rapidly evolving field of chemical risk assessment (CRA).  

The program featured presentations and interactive discussion in breakout sessions dedicated to five core modelling areas:  

• Physiologically Based Kinetic (PBK) Models: Addressing both toxicokinetic and pharmacokinetic models.
• Quantitative Structure-Activity Relationships (QSAR): Including modern Machine Learning and Artificial Neural Network approaches.
• Artificial Intelligence (AI) in CRA: Exploring non-QSAR applications for CRA.
• Exposure models: Highlighting advances in exposure modelling.

Making our data and models FAIR: Discussing challenges and opportunities for organising and integrating data and models to support CRA.

Ahead of the main meeting, a half-day QsarDB.org ↗ workshop introduced open and FAIR approaches for presenting Machine Learning and Artificial Intelligence models.

During the meeting, partners from each modelling area shared their achievements to date, analyses of existing knowledge, and progresses on organising data and models. Participants also discussed future plans, including new PARC initiatives such as the READYAI project.  

This meeting reinforced the value of collaboration across PARC, setting the stage for even more impactful advances in chemical risk assessment. We hope to revisit this fruitful discussion in two years’ time - or even sooner - to further strengthen our collaboration, highlighted Uko Maran from Tartu University.

This was the first time that modelers from all different work packages met together in one event. The presentations and joint discussions helped everyone gain a broader understanding of ongoing efforts in Machine Learning and Artificial Intelligence applications for chemical hazard frameworks, paving the way for greater synergy in future developments.