engleski

Prediction of biodegradability and toxicity of aromatics in water using QSAR modeling

The study was aimed at developing models for predicting the biodegradability and toxicity of aromatic water pollutants. For that purpose, 32 single-benzene ring compounds, with different type, number and position of substituents, were used. Their biodegradability was estimated according to the ratio of the determined biochemical and chemical oxygen demand values. Their activity towards Vibrio fischeri was measured and expressed in toxicity units. The models were derived using quantitative structure- activity relationship (QSAR) principles and tools. The structural features of the studied compounds were described by molecular descriptors derived through DRAGON 3.0 software and by semiempirical MNDO and AM1 methods using their optimized molecular structures. Upon derivation of the models and calibration and subsequent testing on the training and the test set, respectively, 4- variable models were selected as the most predictive in both cases of the activities studied. The most contributing structural feature influencing biodegradability of the studied compounds is molecular mass. The toxicity strongly depends on the symmetry and shape of the molecular structure, while weighting schemes of descriptors included suggest the influence of electrotopological states as well. Both models were verified on two external compounds yielding rather highly accurate prediction of their biodegradability and toxicity.

aromatics ; biodegradability ; toxicity ; structural-relationship modeling ; structure-activity relationship

Ovaj rad je izrađen u sklopu projekta „Modeliranje okolišnih aspekata napredne obrade voda za razgradnju prioritetnih onečišćivala“ Hrvatske zaklade za znanost na Fakultetu kemijskog inženjerstva i tehnologije Sveučilišta u Zagrebu.