Bioconcentration modelling of alcohol ethoxylates by quantitative structure activity relationship approach: a first look

Alcohol ethoxylates (AEs) are a class of nonionic surfactants. This study overviewed the environmental health effects and quantitative structure activity relationships generated for bioconcentration factors of seventeen alcohol ethoxylates, which are currently in commercial use as household detergents. The X-data matrix consisted of 560 molecular descriptors which was calculated by the DRAGON® molecular modelling environment. The logarithms of bioconcentration factors calculated by EPI® toxicology estimation suite were used as the response factor. Out of two quantitative structure activity relationships generated, one exhibited a model fit of 0.95 and a power of prediction of 0.42. The second was superior in terms of model fit, which was 0.92 and a power of prediction of 0.7. The predicted bioconcentration values exhibited a minimum percentage error of 10 % and a maximum of 37 %. Prediction accuracy became better with increasing bioconcentration factor. A convincing relationship between bioconcentration and calculated molecular descriptors for alcohol ethoxylates was obtained, hence the capability of quantitative structure activity relationship approach for modelling the environmental behaviour of AEs at a fully empirical level was demonstrated. Constructing a quantitative structure activity relationship having a realistic predictive power over a variety of commercial AEs may be challenging, but with the use of finely tuned chemical descriptors and better modelling tools it could be possible to accurately and rapidly predict toxicities as well as the environmental behaviour of AEs.