Spectral prediction of the botanical identity of organ powders from Zanthoxylum zanthoxyloides (Lam.) Zepernick and Timler

---

Résumé

Zanthoxylum zanthoxyloides is a medicinal plant employed in the production of phytomedicines in Burkina Faso, including FACA®, which is used for the treatment of sickle cell disease.
A chemometric spectral-prediction approach was applied to classify Z. zanthoxyloides organ powders according to organ type, including leaves, stem bark, and roots.
The characteristic FTIR spectral profile of Z. zanthoxyloides powders revealed spectral bands of interest between 800 and 1750 cm−1, enabling discrimination among the analyzed powders. A Partial Least Squares–Discriminant Analysis (PLS-DA) model was constructed using spectral data from 75 powder samples representing three plant organs, collected from three sites in Burkina Faso.
The PLS-DA model, validated through ANOVA and permutation tests, demonstrated excellent performance, achieving 93.33% overall classification accuracy and an ROC-AUC exceeding 0.90 across the three organ classes. This statistical model enabled the determination of the botanical affiliation of 33 new plant powders by accurately assigning them to their respective organ classes of Z. zanthoxyloides.
Chemical characterization of powders using their FTIR spectral profiles represents a green analytical approach contributing to the standardization of botanical raw materials. Integrating chemometric and multivariate data-analysis tools into raw-material quality-control routines will enhance the safety and reliability of locally produced phytomedicines.

Mots-clés

Chemometrics, FTIR, PLS-DA, Raw materials standardization, Zanthoxylum zanthoxyloides