Performance evaluation of drug synergy datasets using computational intelligence approaches

Drug synergy has become a promising approach for malignant diseases. This approach can increase therapeutic efficacy, reduce toxicity, and overcome drug resistance compared with single-drug administrations. Thus, it has attracted a lot of interest from researchers and pharmaceutical companies. It is critical for scholars interested in this field to be aware of relevant datasets to discern their findings and (i) understand how to use them (ii) instead of creating custom datasets, they can speed up their research by utilizing the pre-existing datasets. In this study, the combined activity of anticancer drugs are modelled with computational techniques random forest, linear regressor, decision tree, and adaBoost regressor. The drug pairs screening data from NCI-ALMANAC and Merck is used for training computational techniques. Using the root mean squared error (RMSE), mean absolute percentage error (MAPE), Pearson's correlation coefficient, and coefficient of determination (R2) for various computational intelligence techniques were analyzed. All these measures are used to understand the behaviour of the computational techniques and the characteristics of the datasets. The performance metric RMSE and MAPE focus on synergy score, whereas Pearson, R2 focuses on the feature set. The R2 metric is a good choice for the selection of computational technique. The analysis of computational techniques presented in this paper shows the promising direction in the study related to drug synergy using existing datasets which have been created by reliable and authenticated national cancer institute (NCI) and O'Neil study. By using computational techniques, researchers can shorten the delivery time for new drugs thereby saving money as well as time.

Automated summary

Drug synergy is a promising approach for malignant diseases, which can improve therapeutic efficacy, reduce toxicity, and overcome drug resistance. This study uses computational techniques to model the combined activity of anticancer drugs and analyzes their performance using various metrics. By utilizing existing datasets and computational techniques, researchers can accelerate the development of new drugs, saving both time and money.