In Silico Analysis of Bioactive Compounds from Medicinal Plants: Molecular Docking, ADMET Profiling, and Molecular Dynamics Validation

Abstract

Medicinal plants represent a vast and largely underexplored repository of pharmacologically active molecules. Traditional medicine systems across Africa, Asia, and Latin America have long leveraged plant-derived bioactive compounds for the treatment of infectious, metabolic, and neoplastic diseases. However, the systematic scientific validation of these compounds has been constrained by high costs and time requirements of conventional drug discovery. In silico approaches—encompassing molecular docking, ADMET prediction, and molecular dynamics (MD) simulation—offer computationally efficient platforms for the preliminary screening and prioritization of candidate phytochemicals before experimental validation. This study aimed to perform a comprehensive in silico analysis of 30 bioactive compounds from 10 ethnopharmacologically important medicinal plants against 10 disease-relevant protein targets, evaluate their drug-likeness using Lipinski's Rule of Five and ADMET profiling, and validate top hits through molecular dynamics simulation. Bioactive compounds were retrieved from PubChem and the Universal Natural Products Database (UNPD). Protein crystal structures were obtained from the RCSB Protein Data Bank (PDB). Molecular docking was performed using AutoDock Vina 1.2. ADMET properties were predicted using SwissADME, pkCSM, and ProTox-II. Drug-likeness was assessed via Lipinski's Rule of Five (Ro5) and Veber criteria. Top-ranked compounds underwent 50 ns MD simulation in GROMACS 2022 using the CHARMM36 force field. Binding free energies were estimated using the MM-PBSA approach. Of the 30 compounds screened, 22 satisfied Lipinski's Ro5 criteria. Molecular docking identified EGCG (ΔG = −9.4 kcal/mol vs. α-glucosidase), Withaferin A (ΔG = −9.1 kcal/mol vs. NF-κB p65), and Quercetin (ΔG = −8.6 kcal/mol vs. EGFR) as top binding candidates. ADMET profiling revealed favourable pharmacokinetic profiles for Withaferin A, Andrographolide, and Curcumin. MD simulation confirmed the stability of the Withaferin A–NF-κB complex (mean RMSD = 0.15 nm over 50 ns) with MM-PBSA binding free energy of −42.3 ± 3.1 kJ/mol. This study demonstrates that Withaferin A, Andrographolide, and Curcumin are strong candidate phytochemical leads warranting further in vitro and in vivo validation. The integrated in silico pipeline employed offers a reproducible, cost-effective framework for early-phase drug discovery from medicinal plant sources, with significant translational relevance for resource-constrained research environments in Nigeria and sub-Saharan Africa.