Evaluation of cisplatin efficiency as a chemotherapeutic drug based on neural networks optimized by genetic algorithm

Cisplatin is an active drug against many types of cancers; its effect appears through genetic toxicity which caused by interaction with the DNA of the cell. The gene expressions prediction of the patients is a very vital process in estimating the drug response. In this paper, we proposed an optimized Neural Networks (NNs) by Genetic Algorithm (GA) for evaluation of cisplatin efficiency as a chemotherapeutic drug. The proposed approach minimizes the error between the actual and the predicted genes until reaching the minimum Mean Square Error (MSE) of NNs which accordingly, improve the prediction accuracy. We used a public dataset (divided into five sub-datasets), where that data demonstrated the genotoxicity of different chemicals like cisplatin, sodium, chloride, and taxol. It was used only cisplatin as an indicator of DNA damage. The prediction accuracy of the optimized NNs by GA was high as lower MSE achieved in all sub-datasets.