Predicting the Anticancer Activity of 2-alkoxycarbonylallyl Esters against MDA-MB-231 Breast Cancer - QSAR, Machine Learning and Molecular Docking

Background: The continuous increase in mortality of breast cancer and other forms of cancer due to the failure of current drugs, resistance, and associated side effects calls for the development of novel and potent drug candidates. Methods: In this study, we used the QSAR and extreme learning machine models in predicting the bioactivities of some 2-alkoxycarbonylallyl esters as potential drug candidates against MDA-MB-231 breast cancer. The lead candidates were docked at the active site of a carbonic anhydrase target. Results: The QSAR model of choice satisfied the recommended values and was statistically significant. The R²pred (0.6572) was credence to the predictability of the model. The extreme learning machine ELM-Sig model showed excellent performance superiority over other models against MDAMB- 231 breast cancer. Compound 22 with a docking score of 4.67 kcal mol^-1 displayed better inhibition of the carbonic anhydrase protein, interacting through its carbonyl bonds. Conclusion: The extreme learning machine’s ELM-Sig model showed excellent performance superiority over other models and should be exploited in the search for novel anticancer drugs.