Which chemicals can be excluded from further assessment? (“true negatives”)
Order all chemicals on the Risk Score assuming that the model underestimates the PEC by 2 orders of magnitude (“Score (PEC*100)”).
Filter out all simulations which are not correct.
This provides 1220 chemicals with a risk score of 0.
These chemicals are not expected to cause any problem, regardless of the model accuracy.
Which chemicals can be classified as “high priority”? (“true positives”)
Order all chemicals on the Risk Score assuming that the model overestimates the PEC by 2 orders of magnitude (“Score (PEC/100)”).
Filter out all simulations which are not correct.
This provides only 1 chemical with a risk score higher than 0 (CAS 541-02-6, decamethylcyclopentasiloxane).
By inspection of the FoE, we conclude that this chemical exceeds the PNEC at at least 12% of sites, regardless of the model accuracy, and can be prioritised anyhow.
How to rank the remaining chemicals?
Order all chemicals:
first on the Risk Score assuming that the model correctly simulates the PEC (“Score (PEC)”);
next on the Risk Score assuming that the model underestimates the PEC by a factor of 100 (“Score (PEC*100)”).
Filter out all simulations which are not correct.
The “Score (PEC)” now provides the ranking, while the numbers in “Score (PEC/100)” and “Score (PEC*100)” show the expected range of the score if the model inaccuracy is considered.