PREMIER partners publish new scientific paper: Systematic Handling of Environmental Fate Data for Model Development─Illustrated for the Case of Biodegradation Half-Life Data

To assess the potential risk of a pesticide, pharmaceutical, or any other chemical in the environment, we need to know how fast it can be degraded. Models to estimate environmental half-lives of such micropollutants rely on experimental evidence from biodegradation tests of other compounds. However, the quality of available data is highly heterogeneous, which makes it difficult build models that are predictive with acceptable certainty.

In this work, we propose a statistical approach to systematically characterize the experimental variability and uncertainty of biodegradation data, and we apply it to a data set of biotransformation half-lives for pesticides in soil. This enhanced data set can be used in future to build models using artificial intelligence, that not only predict biotransformation half-lives, but also the statistical uncertainty of the predictions. By providing a statistically sound approach to describe available knowledge, we contribute to the improvement of computational tools to assess the environmental risk of chemicals in the environment. Such tools can be used by regulators in chemical risk assessment and by industry for the design of environmentally benign and readily biodegradable chemicals for the future.

See the article here.