Creation of Quantitative Feature Toxicity Relationship Models for Cytotoxicity of Cadmium Containing Quantum Dots Towards HEK Cells Using QuasiSMILES

Abstract

The rising possibilities of quantum dots in biological fields have raised significant concerns about their toxicological influence, and there is a requirement of methods for prediction of toxic quantum dots. In the current research work, quantitative feature toxicity relationship models for HEK cells cytotoxicity of forty one CdSe/CdTe/CdZnS core quantum dots as a function of their size, surface ligand, ligand chemical, charge, surface modification, assay type and quantum dot concentration have been developed. The models are constructed by using quasiSMILES representation symbolizing the experimental conditions and molecular architecture of the quantum dots. The index of ideality of correlation helps in the building of eight statistically significant, robust and predictive models using Monte Carlo optimization. The results are used successfully for extraction of features causing an increase and decrease of cytotoxicity of quantum dots.