Abstract / Summary
SARS CoV 2 main protease (Mpro) is an essential cysteine protease and a validated antiviral target. This study investigated GC MS identified phytochemicals from Xeromphis nilotica stem bark as potential Mpro inhibitory scaffolds using an integrated computational workflow. GC MS profiling and structural curation produced 158 valid phytochemicals for screening. Machine learning based QSAR modelling was used for initial compound prioritisation, followed by SwissADME assessment, molecular docking, 500 ns molecular dynamics simulation, MM GBSA binding energy analysis, and density functional theory calculations. LightGBM showed the best QSAR performance, with validation R² = 0.640 and external R² = Q² = 0.492. Because the model did not satisfy the predefined Tropsha validation criteria, it was used only for exploratory ranking rather than quantitative potency prediction. SwissADME retained seven drug like candidates, and six compounds showed favourable binding within the Mpro catalytic pocket, although all had weaker docking affinity than nirmatrelvir. Among the simulated phytochemicals, CID_146102 showed the more favourable MM GBSA binding free energy of −36.36 ± 0.46 kcal/mol compared with −30.64 ± 0.49 kcal/mol for CID_415569, together with a more favourable dynamic stability profile during the 500 ns molecular dynamics simulation. Its docking score of −6.56 kcal/mol provided complementary structural evidence of active site occupancy. Nirmatrelvir remained more favourable, with an MM GBSA binding free energy of −46.67 ± 0.52 kcal/mol. DFT analysis further supported the favourable electronic properties of CID_146102. Overall, CID_146102 emerged as the most promising X. nilotica derived Mpro lead like scaffold identified in this study, but its predicted activity remains computational and requires confirmation through enzymatic inhibition, cytotoxicity, and cell based antiviral assays .