Integrated In Silico Models for the Prediction of Human Repeated Dose Toxicity of Cosmetics to Optimise Safety

The COSMOS project will address the assessment needs of the cosmetics industry by delivering an integrated suite of computational tools to predict the effects of long-term exposure to cosmetic ingredients in humans, thus reducing the need for repeated dose toxicity testing in animals. To achieve this, individual modules comprising: (new) databases, thresholds of toxicological concern (TTC), in silico toxicology (grouping, read-across and QSAR), in vitro data and physiologically-based pharmacokinetic (PBPK) modelling, will be used to construct flexible workflows for assessing toxicity. The COSMOS project will be informed by the needs of industry through active stakeholder engagement, and will utilise innovative technologies from outside the cosmetics area. New databases for TTC and modelling studies will be created by harvesting US FDA legacy data for cosmetics. New and current databases will be combined to advance the state-of-the-art providing a transparent, freely-available public resource. Grouping and read-across will be achieved mechanistically and on the basis of structural similarity. The effective dose-response at the target organ will be estimated from in vitro effects and PBPK models, establishing an alternative (non-animal) basis for risk assessment. All models and Workflows developed will be transparent, fully documented and open access. The partners include world leaders committed to donating software and algorithms to support the open architecture. Data and models will be integratable with existing systems . Workflows will be automated through the KNIME software.