Supervised\nScreening of EGFR Inhibitors Validated\nthrough Computational Structural Biology Approaches

One of the prominent challenges in breast cancer (BC)\ntreatment\nis human epidermal growth factor receptor (EGFR) overexpression, which\nfacilitates tumor proliferation and presents a viable target for anticancer\ntherapies. This study integrates multiomics data to pinpoint promising\ntherapeutic compounds and employs a machine learning (ML)-based similarity\nsearch to identify effective alternatives. We used BC cell line data\nfrom the Cancer Cell Line Encyclopedia (CCLE) and Genomics of Drug\nSensitivity in Cancer (GDSC) databases and single-cell RNA sequencing\n(scRNA-seq) information that established afatinib as an efficacious\ncandidate demonstrating superior IC<sub>50</sub> values. Next, ML\nmodels, including support vector machine (SVM), artificial neural\nnetworks (ANN), and random forest (RF), were trained on ChEMBL data\nto classify compounds with similar activity to the reference drug\nas active or inactive. The promising candidates underwent computational\nstructural biology assessments for their molecular interactions and\nconformational dynamics. Our findings indicate that compounds ChEMBL233324,\nChEMBL233325, ChEMBL234580, and ChEMBL372692 exhibit potent repressive\naction against EGFR, underscoring their potential as active antibreast\ncancer agents.