Recent emphasis on the development of safe-and-sustainable-by-design chemicals highlights the need for methods facilitating the early assessment of persistence. Activated sludge experiments have been proposed as a time- and resource-efficient way to predict half-lives in simulation studies. Here, this persistence "read-across" approach was developed to be more broadly and robustly applicable. We evaluated 21 previously used reference plant protection products (PPPs) for their broader applicability in calibrating regression and classification models for predicting half-lives in soil (DT50) and water-sediment systems (DT50) based on their half-life in sludge and the organic carbon-water partition coefficient as predictors. The calibrated regression models showed satisfactory predictions of DT50 for another 22 test PPPs. Performance was less satisfying for the prediction of DT50 for 46 active pharmaceutical ingredients (APIs), suggesting a need for expanding the set of calibration substances and more experimental values. The classification models mostly correctly classified persistent and non-persistent test compounds for both PPPs and APIs, which is relevant for early-stage screening of persistence. Transformation products of the reference compounds in activated sludge samples were consistent with the reported degradation pathways in soil, particularly with respect to major aerobic, enzyme-catalyzed transformation reactions. Overall, "reading across" biotransformation in environmental compartments such as soils or sediments from experiments with activated sludge outperformed three widely used approaches for estimating half-lives and hence has immediate potential to support early assessment of biodegradability when aiming to develop chemicals that are safe and sustainable by design.
Coll et al. (2025) studied this question.