Advances in AI for predicting pharmacological properties of natural medicines

Artificial intelligence (AI), which includes machine learning (ML) and deep learning (DL), has become an important tool in drug development. An increasing number of studies have discovered pharmacologically active compounds in natural products, and AI's high-throughput capabilities have accelerated the drug discovery process. In the development of natural medicines, AI can use existing datasets and experimental data to screen for lead compounds with potential activity and predict disease-associated drug targets. Furthermore, using AI to predict the absorption, distribution, metabolism, excretion, and toxicity (ADMET) properties of lead compounds early in the drug discovery process can save time and money. This review introduces AI applications for predicting the pharmacological properties of natural drugs by outlining model construction principles and recent advances, summarizing key aspects such as feature selection and evaluation metrics, and discussing natural drug development challenges and opportunities. This review focuses on the advances of artificial intelligence (AI), including machine learning (ML) and deep learning (DL), in predicting the pharmacological properties of natural medicines, highlighting its roles in screening potential active compounds, identifying drug targets, and forecasting absorption, distribution, metabolism, excretion, and toxicity (ADMET) properties to accelerate drug discovery. It details various AI models such as LASSO, SVM, XGBoost, GNN, and CNN, along with key aspects like multimodal data integration and reliance on specialized databases for high-quality data support. Despite AI's high-throughput and accurate predictive capabilities, the field faces challenges including incomplete natural product databases, poor model interpretability, and insufficient integration with clinical data. The review emphasizes the need for standardized databases, enhanced model explainability, and more clinical validation to bridge preclinical research and real-world applications. Ultimately, AI is positioned as a transformative tool for unlocking the potential of natural medicines, with future efforts focusing on comprehensive data networks and translational research to improve drug development efficiency and safety.