Aromatic Compounds Aquatic Toxicity QSAR Using Molecular Quantum Similarity Measures

Abstract
Molecular Quantum Similarity is discussed to be a satisfactory chemical tool to develop QSARs for the toxicity of aromatic narcotic pollutants to fish Poecilia reticulata. The method is presented as a possible alternative to the use of classical 2D QSAR descriptors, such as logP, Hammett's sigma, or HOMO and LUMO energies, and it is based upon the information extracted from the quantitative similarity measures between the overall molecular pairs of the studied set. A systematic study of medium size homogeneous sets is performed, and finally an analysis of the full set is carried out. In all cases, satisfactory prediction models are achieved using few parameters. The stability and reliability of QSARs are tested by cross-validation, randomisation and real prediction.

This publication has 26 references indexed in Scilit: