HOME     SCHEDULE     AUTHOR INDEX     SUBJECT INDEX         

PARENT SESSION

PW03 QSAR
Exhibit Hall
8:00 AM - Wednesday

(PW029) The limitations of commonly used ecotoxicity QSARs – experiments in model validation.

Salvito, D1, 1 Research Institute for Fragrnace Materials, Woodcliff Lake, NJ, USA

ABSTRACT- QSARs are becoming increasingly important tools to screen and, in some cases, assess the fate and effects of large groups of organic compounds. Tens of thousands of organic compounds are in the process of being screened for hazard classification under several international regulatory initiatives. A reliance on QSARs is likely; however, the limitations of the available models may be underestimated and unrecognized. Two commonly used QSAR models are the Könemann equation and ECOSAR, software available as part of part of Syracuse Research Corporation′s EPIWIN Suite. ECOSAR uses a number of log Kow based QSARs to estimate the ecotoxicity of organic compounds for several structural classes often resulting in ecotoxicity estimates of a variety of endpoints. In an effort to demonstrate the validity of the outputs of these models, a set of 44 fragrance materials was selected to compare the models′ output against measured data for lethality (96 hour-LC50) for Pimephales promelas (Fathead minnow). A method was developed to separate narcosis compounds from non-narcosis compounds. The two methods used to separate the compounds were the determination of the mode of action using a classification system reported by Verhaar et al. and the development of a log Kow based toxicity relationship based on the work of DiToro et al. Twenty-four fragrance materials, using both methods, were categorized as definite narcosis compounds. The remainder did not fit both definitions of narcosis compounds. The correlations, using either the Könemann equation or the ECOSAR model, were very good (r2 of 0.84 and 0.93, respectively) for the set of definite narcosis compounds and very poor for the remaining fragrance materials (r2 of 0.065 and 0.063, respectively). It is apparent that these models are capable of both underpredicting and overpredicting toxicity.

Key words: hazard assessment, QSAR, ecotoxicity


Internet Services provided by
Allen Press, Inc. | 810 E. 10th St. | Lawrence, Kansas 66044 USA
e-mail assystant-helpdesk@allenpress.com | Web www.allenpress.com
All content is Copyright © 2004 SETAC