PARENT SESSION

1F - QSAR
Poster Hall
8:30 AM - Tuesday, 29 April 2003
Chair: Schüürmann, G.¹, ¹
Co-chair: Verhaar, H.J.M.², Cronin, M.³, ^{2 3}

(TUP/45) Prediction of acute toxicity in fathead minnow based on different QSAR approaches.

Colombo, Andrea¹, Fattore, Elena ¹, Piclin, Nadège^{1, 2}, Roncaglioni, Alessandra¹, Benfenati, Emilio¹, ¹ Istituto di Ricerche Farmacologiche, Milan, Italy² Biochemics Consulting, Orléans, France

ABSTRACT- QSAR is an useful technique to predict the risk of a substance from its chemical structure when experimental data are not available. This study provides an investigation and comparison between three different strategies of QSAR modelling for aquatic toxicity based on: i) sub-models for different chemical classes; ii) sub-models for different Modes Of Action (MOAs); iii) all substances without any a priori knowledge. We considered 568 chemicals, belonging to fifteen chemical classes, and characterised by the experimental assignment to a specific MOA; the 96-hours Lethal median Concentration (LC₅₀ 96h) for fathead minnow (Pimephales promelas) was investigated (Predicting modes of action from chemical structure: acute toxicity in the fathead minnow, Russom C.L. et al., Environ. Toxicol. Chem., 16, 948, 1997). About 190 chemical descriptors were calculated with different software in order to obtain different categories of input variables (topological, geometrical, constitutional, electrostatic, physic-chemical and quantum chemical). Partial Least Square (PLS) method was utilised to predict LC₅₀ 96h, and internal and external validation procedures were employed to validate the different models. Best predictive models (R²_cv > 0,9) were obtained with the subsets of chemicals, based on both the assignment to the same chemical class or MOA, and using not more than four chemical descriptors. On the other hand, models developed with the whole data set showed a R²_cv minor of 0,7 using five descriptors. These preliminary results involve the issue of the development of appropriate tools for the a priori assignment to a specific QSAR model in order to obtain a capable and predictive system.

Key words: Aquatic Toxicity, QSAR, Fathead Minnow