Topological quantum similarity measures: applications in QSAR

Abstract

A novel method for computing new descriptors to construct Quantitative Structure–Toxicity Relationships is presented. First, a brief review on the classical graph theory is presented and, then, the link with molecular similarity is drawn. In the applications section, molecular topological indices are calculated using the interatomic Molecular Quantum Similarity Measure with a Coulomb weight operator. The use of similarity matrices instead of classical topological ones has been adopted according to the connection between molecular topology and the general theory of Quantum Similarity. Afterwards, the molecular descriptors, which include the structural information necessary to properly describe the system, are employed to derive numerical correlation with toxicities. The QSAR model is built using a multilineal regression technique. Finally, some application examples are presented, including polycyclic aromatic hydrocarbons and aquatic toxicants, demonstrating the applicability of the exposed methodology.