Integrating HPGPC and FT-IR fingerprints with multi-block data fusion for origin authentication and quality assessment of Glycyrrhiza polysaccharides

Glycyrrhizae polysaccharides (GPs) are key bioactive components with immunomodulatory and antioxidant properties, yet their structural complexity and origin-dependent variability pose challenges for quality control. This study established a comprehensive quality evaluation system for GPs from three official Glycyrrhiza species ( G. uralensis , G. inflata , and G. glabra ) by integrating HPGPC and FT-IR fingerprinting with multivariate chemometric and data fusion strategies. HPGPC revealed six common peaks with distinct molecular weight distributions, while FT-IR spectra indicated species-specific differences in functional groups and glycosidic linkages. Chemometric models (PCA, HCA, PLS-DA, BP-ANN, SIMCA, CP-ANN) effectively discriminated GPs by botanical origin. Notably, mid-level data fusion (MLDF) of HPGPC and FT-IR data, combined with PLS-DA, achieved 100 % classification accuracy, outperforming low-level fusion and single-technique models. This multi-modal approach provides a robust and efficient framework for authenticating Glycyrrhiza polysaccharides, facilitating quality standardization and origin traceability in herbal medicine.