What is it about?
The paper looks at a few predictive models for acute aquatic toxicity and their methodologies. The model performance is assessed based on a common data set external to all training sets. Regulatory-relevant toxicity categories are included in assessment. Finally, we track differences in predicted values to the estimation methods, whenever possible.
Why is it important?
It is always important to validate models with new data. It is equally vital to differentiate modeling methods that yield best results. Thus, the analysis presented in the paper aims to give some guidance on the effectiveness of different model methodologies. We hope that the information will help other scientist build more robust in silico tools by considering properties used in the model and potential modes of chemical toxicity.
The following have contributed to this page: Fjodor Melnikov
In partnership with: