Machine learning-based screening of MCF-7 human breast cancer cells and molecular docking analysis of essential oils from Ocimum basilicum against breast cancer

Abstract

A machine learning algorithm-based model is a powerful tool for discovering candidate compounds on the breast cancer cell line MCF-7. Using the Lazy Predict Python package, the “Random Forest” algorithm indicates that the highest accuracy is 83.57% in discovering potential compounds. Essential oils from the leaves and stems of Ocimum basilicum, as identified by GC−MS analysis, are selected as a test dataset. Among them, eight essential oils, including alpha-pinene, trans-beta-ocimene, estragole, alpha-cubebene, gamma-muurolene, delta-cadinol, gamma-cadinene, and beta-ocimene potentially exhibit activity against MCF-7. The anticancer mechanisms of these essential oils are analyzed using molecular docking simulation based on the structure-activity relationship between these candidates and the two protein targets, BRCA1 and BRCA2. This study shows that our model can potentially screen bioactive compounds targeting breast cancer cell line MCF-7 and offer the basis for further research into substances derived from Ocimum basilicum that can potentially be utilized as a novel treatment for breast cancer.