From target fishing to AI and single-cell multi-omics: an integrated framework for natural product target research

Natural products are important sources of therapeutic agents, yet their structural diversity, polypharmacology, and context-dependent effects complicate target discovery. This review presents an integrated, decision-oriented framework linking experimental target fishing, orthogonal target-engagement and functional validation, single-cell and spatial multi-omics, and artificial intelligence (AI)-assisted prioritization. We compare representative label-based and label-free target-fishing strategies according to their biological applicability, evidential strength, limitations, and validation requirements. We also outline practical approaches for resolving drug-responsive cell states and tissue niches using single-cell and spatial technologies, and summarize AI methods for compound-target prediction, graph- and knowledge-based reasoning, perturbation modeling, and multimodal data integration. Particular attention is given to data quality, applicability domains, interpretability, and prospective validation. Finally, we propose an evidence-informed closed-loop workflow in which computational and omics-derived hypotheses are tested through direct-binding, cellular-engagement, genetic, pharmacological, and spatially resolved experiments. This framework aims to improve the rigor and efficiency of natural product target discovery and mechanism elucidation.