Ecotoxicological Modeling of Organic Chemicals for Their Acute Toxicity in Algae Using Classification and Toxicophore-Based Approaches
Kabiruddin Khan, Probir Kumar Ojha · International Journal of Quantitative Structure-Property Relationships · 2020
To study the relationship between toxicity of organic chemicals (OCs) with their structural and physicochemical features, the authors have developed a linear discriminant analysis (LDA) model for the classification of organic chemicals based on their acute observed toxicity in algae. This is done by employing a sufficiently large dataset of 352 chemicals following the strict Organization for Economic Cooperation and Development (OECD) guidelines for the quantitative structure-activity relationship (QSAR) validation. Additionally, 3D toxicophore models were generated to explore for the presence of common features contributing to the toxicity making chemicals a major concern for the future. Both of the models were rigorously validated following stringent validation criteria such as Wilks' λ statistic, canonical index (Rc), squared Mahalanobis distance, and chi-squared. Finally, a confusion matrix was employed to check for the quality of classification/prediction obtained for the LDA and pharmacophore models both in the training and the test sets.