Conceptos Básicos
Efficient identification of novel inhibitor molecules targeting EGFR, HER2, Estrogen, Progesterone, and NF-κB receptors for breast cancer treatment through machine learning-based virtual screening and molecular docking.
Resumen
This study aimed to develop efficient computational methods for virtual screening of chemical compounds against key breast cancer targets, including EGFR, HER2, Estrogen Receptor (ER), Progesterone Receptor (PR), and Nuclear Factor-kappa B (NF-κB).
The researchers first constructed binary and multi-class classification models using various machine learning techniques, including Genetic Algorithm (GA), Support Vector Machine (SVM), Random Forest (RF), and Quadratic Discriminant Analysis (QDA). The GA-SVM-SVM:GA-SVM-SVM pipeline emerged as the most effective, achieving an accuracy of 0.74 and an AUC of 0.94 for virtual screening.
This pipeline was then used to screen the BindingDB database, identifying 4454, 803, 438, and 378 new inhibitor molecules for EGFR+HER2, ER, NF-κB, and PR, respectively, with over 90% precision in both active/inactive and target prediction.
The binding energies of the selected molecules were evaluated through molecular docking, revealing a range of -15 to -5 kcal/mol, which is considered suitable for inhibiting the target receptors. Further prioritization of the new molecules was performed based on established medicinal chemistry rules and parameters, such as Lipinski, Pfizer, GSK, and golden triangle rules, as well as QED, SAscore, and MCE-18.
The study provides a comprehensive computational framework for efficient virtual screening and identification of novel breast cancer inhibitors, which can accelerate the early-stage drug discovery process.
Estadísticas
The binding energy range for the selected ligands against their respective targets was calculated to be between -15 and -5 kcal/mol.
Citas
"The GA-SVM-SVM:GA-SVM-SVM model was selected with an accuracy of 0.74, an F1-score of 0.73, and an AUC of 0.94 for virtual screening of ligands from the BindingDB database."
"This pipeline successfully identified 4454, 803, 438, and 378 ligands with over 90% precision in both active/inactive and target prediction for the classes of EGFR+HER2, ER, NF-κB, and PR, respectively, from the BindingDB database."