Search for a command to run...
The research investigates how allicin from garlic interacts with flavonoids to create synergistic antioxidant effects, which shows promise for developing new nutraceuticals. The process of testing different allicin-flavonoid pairs through experimental methods proves too expensive for practical use. The research establishes a predictive machine learning system to find the most effective allicin-flavonoid combinations with synergistic effects at high speed. The research used a computational method that combined Density Functional Theory (DFT) with Quantitative Structure-Activity Relationship (QSAR) modeling. The research team built a diverse flavonoid dataset and performed molecular descriptor calculations. The B3LYP/6-311+G(d,p) DFT method calculated precise Bond Dissociation Enthalpy (<tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\triangle \text{BDE}$</tex>) changes for a selected set of compounds when they formed complexes with allicin. The Random Forest Regressor model received <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\triangle \text{BDE}$</tex> data as its target variable to establish a relationship between molecular structures and their synergistic effects. The trained model achieved outstanding predictive accuracy, as indicated by an <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathrm{R}^{2}$</tex> score of 0.89 when evaluated on a separate test dataset. The number of hydroxyl groups (nOH) emerged as the primary factor, explaining more than half of the predictive power, according to feature importance analysis, followed by topological polar surface area (TPSA). The model demonstrates that H-bond donation ability plays a vital role in stabilizing radicals. The research created a dependable ML-QSAR system that enables scientists to design antioxidant combinations through rational methods. The model serves as an efficient high-speed screening system for large virtual flavonoid libraries, enabling the selection of optimal candidates for experimental verification, thereby accelerating the development of therapeutic and nutraceutical agents against oxidative stress.