AI-Based Prediction of Pharmacological Activities of Plant-Derived Natural Products: Advances, Applications, Challenges, and Future Perspectives

Abstract

Plant-derived natural products remain one of the most important sources of bioactive compounds for drug discovery owing to their remarkable chemical diversity and broad spectrum of pharmacological activities. However, conventional approaches for identifying therapeutically relevant phytochemicals are labor-intensive, time-consuming, and associated with high costs and low success rates during lead optimization. Recent advances in artificial intelligence (AI) have transformed natural product research by enabling rapid prediction of pharmacological activities through the integration of machine learning, deep learning, and cheminformatics approaches. AI algorithms can analyze large-scale chemical and biological datasets, identify structure–activity relationships, predict molecular targets, estimate pharmacokinetic and toxicity profiles, and prioritize promising compounds for experimental validation. The availability of publicly accessible phytochemical and bioactivity databases has further accelerated the development of robust predictive models for natural product-based drug discovery. This review summarizes recent advances in AI-assisted prediction of pharmacological activities of plant-derived natural products, with particular emphasis on machine learning algorithms, deep learning architectures, molecular representations, and publicly available databases used for predictive modeling. Furthermore, the review discusses recent therapeutic applications across major disease areas, including cancer, infectious diseases, inflammatory disorders, diabetes, and neurodegenerative diseases, while highlighting current challenges, methodological limitations, and emerging opportunities for integrating explainable AI and multi-omics technologies into phytopharmaceutical research. Collectively, AI-driven computational strategies have the potential to accelerate natural product drug discovery, improve prediction accuracy, reduce experimental costs, and facilitate the development of safer and more effective plant-derived therapeutics.