Nano-QSAR models for predicting cytotoxicity of metal oxide nanoparticles (MONPs) to E-coli

作者：Zhou, Zhengwei; Tang, Xinwen; Dai, Wen; Shi, Jingjie; Chen, Haiqun^*

来源：Canadian Journal of Chemistry, 2017, 95(8): 863-866.

DOI：10.1139/cjc-2017-0172

• quantitative structure-activity relationship (QSAR)
• metal oxide nanoparticles (MONPs)
• cytotoxicity
• multiple linear regression (MLR)
• support vector machine (SVM)

Nanotechnology has been applied to many aspects of human life. Meanwhile, concerns regarding the toxicity of engineered nanomaterials to the environment have also been growing. Herein, an economic and convenient approach based on quantitative structure-activity relationship for nanomaterials (nano-QSAR) was proposed to evaluate the cytotoxicity of metal oxide nanoparticles (MONPs) to E. coli. Six molecular descriptors of 17 MONPs were selected and calculated using Gaussian98 software and DFT-B3LYP method on the LANL2DZ basis set. Two multivariable models, linear and nonlinear, were built based on the calculated molecular descriptors using multiple linear regression (MLR) and support vector machine (SVM) methods, respectively. Results demonstrated that both models presented high reliability, good predictive performance, and fine generalization ability, with all R-2 values greater than 0.84. It was also revealed that the lowest unoccupied molecular orbital (LUMO) and molar heat capacity (C-p) were the two key descriptors influencing the cytotoxicity of MONPs.

• 出版日期2017-8
• 单位常州大学