Bioprospecting 4.0 of Tropical Medicinal Plants: Integration of Metabolomics, Ethnobotanical Databases and Artificial Intelligence for the Robust Identification of New Bioactive Molecules.

---

Résumé

Tropical medicinal plants constitute an important source of structurally diverse metabolites with potential
value for pharmaceutical, nutraceutical and biotechnological research. Conventional natural-product
discovery pipelines, however, remain limited by slow metabolite identification, fragmented ethnobotanical
information and restricted translation from screening to validated candidates. This review examines
Bioprospecting 4.0 as an integrated framework that combines untargeted metabolomics, ethnobotanical
databases, artificial intelligence, cheminformatics and systems biology for the prioritisation of bioactive
molecules from tropical medicinal plants. Particular attention is given to LC-MS/MS-based metabolomics,
molecular networking, NMR-supported structural elucidation, knowledge graphs and AI-assisted prediction of
biological targets and ADMET properties. The review also considers the role of explainable artificial
intelligence, uncertainty quantification and active learning in improving interpretability, reproducibility and
confidence in candidate selection. In addition, the manuscript discusses multi-omics integration, systems
pharmacology, synthetic biology and translational validation as complementary components of nextgeneration
bioprospecting. Ethical governance, FAIR data principles, biodiversity conservation and Access
and Benefit-Sharing frameworks are addressed as essential requirements for responsible research involving
genetic resources and associated traditional knowledge. Overall, the review highlights that Bioprospecting 4.0
can support a more systematic, transparent and ethically grounded exploration of tropical plant
chemodiversity, while recognising the continuing challenges of data fragmentation, limited infrastructure,
underrepresented biodiversity and predictive uncertainty.

Mots-clés

Bioprospecting 4.0; tropical medicinal plants; untargeted metabolomics; LC-MS/MS; ethnobotanical databases; artificial intelligence; explainable AI; knowledge graphs; ADMET prediction; naturalproduct discovery; FAIR data; biodiversity conservation.