Pesticides Research, 94 – Quantitative Structure-Activity Relationships (QSAR) and Pesticides

Quantitative Structure-Activity Relationships (QSAR) and Pesticides

Summary

In environmental risk assessment, information of environmental fate, behaviour and the toxicity of a chemical substance are of basic need. Quantitative Structure-Activity Relationships (QSAR) is a method to derive certain effects or properties of chemical substances in the absence of experimental data.

For pesticides, the data requirements demanded for their authorisation normally means that sufficient data for a risk assessment exist. This is rarely the case for additives, impurities, degradation or transformation products. Most QSAR models are developed from simple industrial chemicals and usually only a few pesticides are included. Especially, new pesticides consist of complicated molecular structures acting with specific modes of action and their physico-chemical properties and ecotoxicological effect concentrations may not be estimated sufficiently close to the correct effect value by QSAR.

The current document presents the general framework in which QSARs can be used within the risk assessment procedure. Furthermore, it presents recommended QSARs for physico-chemical parameters and ecotoxicological effect concentrations and performs an evaluation of their correlation with approximately 400 selected pesticides including salts and esters.

Of physico-chemical properties, the recommended QSARs to estimate the boiling point could not be evaluated due to lack of experimental values. Estimations of the melting point were inaccurate and mainly overestimated by the presented QSAR. The solubility in water was reasonably well correlated with the recommended QSAR and a QSAR based on 322 pesticides was developed. The vapour pressure QSAR based on suggested values from a computer model showed low correlation with the pesticides used. The Henry's Law constant was evaluated by comparing calculated values with QSAR predicted values. A low correlation was observed. The octanol/water partition coefficient Kow derived by structure fragment analysis demonstrated acceptable agreement with the measured values. As the model was a computerised version, a QSAR based on water solubility was suggested. Several QSARs have been developed to derive the adsorption coefficient Koc. Four recommended QSARs were selected for comparison. The correlations between estimated and experimental Koc were acceptable for one of the QSARs. The QSAR for bioaccumulation was in reasonably good agreement with the experimental values for pesticides.

For aquatic toxicity, QSARs have been developed for acute toxicity estimations for fish, daphnia and algae. The correlations between estimated and experimental EC₅₀-values were low for fish, daphnia and algae. New QSARs, based on experimental toxicity values and log Kow, were suggested following a grouping of pesticides into modes of action. It was observed that improvements of the predictability could be obtained in some of the groups. However in other groups, the correlation coefficients were low and they could not be recommended except perhaps for screening procedures.

QSAR should not replace experimental values for pesticides. However, QSARs proved to have a reasonable predictive value and might be usable if no data were available.