INSIGHT OF NATURAL POLYPHENOL COMPOUNDS FROM ANTIMYOMA TRADITONAL PLANTS: VIRTUAL SCREENING, DRUG LIKENESS, MACHINE LEARNING BASED QSAR AND MOLECULAR DYNAMICS

Uterine fibroids are common benign tumors in women, particularly among those African descent, and are often associated with pain, abnormal bleeding and infertility. However, the limitations and side effect of current surgical and hormonal treatments underscore the need for alterative therapeutic approaches. Therefore, this in silico study evaluated the antimyomatous potential of phytocompounds (polyphenols) from five African medicinal plants (Allium cepa, Combretum micranthum, Vitex agnus castus, Zingiber officinale and Elaeis guineensis), targeting estrogen receptor alpha (ERα). Initially, a total of 399 bioactive compounds were retrieved from the literature and online databases, molecular docking (High throughput virtual screening, Standard precision and extra precision with Schrödinger Glide) was performed with 4-hydroxytamoxifen used as the reference antagonist. As a result, twelve compounds including luteolin (-10.766 Kcal/Mol), orobol (-10.971Kcal/Mol), (-)-epigallocatechin (-10.481 Kcal/Mol) and (-)-epicatechin (-10.481Kcal/Mol) exhibited higher binding affinities than 4-hydroxytamoxifen. Moreover, ligand-protein interactions involve multiple hydrogen bonds and hydrophobic contacts with key active site residues. Subsequently, ADMET (Absorption, Distribution, Metabolism, Elimination and Toxicity) profiling predicted favorable pharmacokinetics, with good oral absorption and moderate toxicity. In addition, quantitative structure activity relationship modeling in MOE (Molecular Operating Environment), using decision-tree-based algorithms (HistGB, LightGBM, Random Forest), showed strong predictive performances (R2=0.82, RMSE=0.77, r=0.91). Furthermore, molecular dynamics simulations (100 ns) confirmed the stability of the luteolin-ERα complex, supporting a possible antagonistic effect.