QSAR Modeling of Metal Oxide Nanoparticle Toxicity Based on Atomic Properties and Molecular Topology

Nanomaterials, particularly metal-oxide nanoparticles (NPs), are widely used in various technological applications such as electronics, optoelectronics, pharmacy, medicine, cosmetics, and textiles due to their versatile surface functions. These NPs can enter the human body through various routes and may have unforeseen negative effects, requiring further research into their toxicity. Developing safe and green nanomaterials requires understanding the physicochemical factors that govern their properties, which can be challenging to study experimentally. Thus, in silico methods such as Quantitative Structure-Activity Relationship (QSAR) can help reduce the burden of experimental testing while boosting the accuracy of harmful impact prediction. The use of atomic molecular descriptors in combination with other types of descriptors can improve the accuracy and comprehensiveness of QSAR models for metal-oxide nanoparticles. Comparative analysis demonstrated that models based on simpler molecular descriptors offered improved robustness and computational efficiency compared to more complex descriptors used in other studies.