Machine learning combined with molecular docking to analyze the molecular network of FB1-induced esophageal cancer

Objective This article explores the molecular network of esophageal cancer (ESCC) induced by Fumonisin B1 (FB1). Method We integrated FB1-related genes with those closely associated with esophageal squamous cell carcinoma (ESCC) from two datasets, GSE23400 and GSE53624. Utilizing machine learning algorithms and molecular docking techniques, we conducted a comprehensive analysis of the molecular network linking FB1 to esophageal cancer. Result Through continuous optimization, we identified 157 potential genes associated with FB1-induced ESCC. Among the 128 predictive models constructed using 12 machine learning algorithms, the Lasso + LDA model emerged as the most effective. Based on this analysis, the high-quality core genes identified were CFD, FAP, EPHX2, and PLA2G7. Notably, FAP and PLA2G7 exhibited significant increases, whereas CFD and EPHX2 demonstrated significant decreases. Molecular docking studies indicate that FB1 exhibits a strong binding affinity for these four core proteins. Conclusion This study demonstrates that the potential mechanism through which FB1 contributes to the development of esophageal cancer may involve alterations in specific genes and signaling pathways. Furthermore, molecular docking results confirmed that FB1 exhibits a high binding affinity for the four core proteins: CFD, FAP, EPHX2, and PLA2G7. This research establishes a crucial foundation for further investigation into the carcinogenic mechanisms of esophageal cancer induced by FB1.