Chemometric insights into milk-carrageenan breaking and gel strength

The relationship between the molecular structure and composition of carrageenan and its milk gelation properties has been studied using chemometric and machine learning tools. Carrageenan types—κ-carrageenan and κ/ι-hybrid— were analyzed for their gel and breaking strengths, crucial for their industrial application as food gelling and thickening agents. Four analytical platforms, Fourier-transform infrared spectroscopy (FT-IR), nuclear magnetic resonance (NMR) spectroscopy, size-exclusion chromatography with multi-angle light scattering detection (SEC-MALS), and inductively coupled plasma mass spectrometry (ICP-MS), were employed to characterize the molecular structure and composition aiming to predict milk-carrageenan breaking strength. Both single and multi-block predictive modeling were applied to predict functionality, challenging the conventional approach that relies only on single analytical platforms for prediction. Support Vector Machine (SVM) trained on FT-IR spectra, achieved the most accurate predictions, indicating its potential as an efficient alternative to traditional characterization methods by requiring only measurements directly on the carrageenan powder rather than the laborious functionality testing. In examining multi-block modeling, particularly through Sequential and Orthogonalized PLS (SO-PLS), the study evaluated the added value of incorporating further analytical blocks. While adding SEC-MALS and ICP-MS data did not significantly improve prediction models, their inclusion enriched the causal understanding of carrageenan's structure-function relationship.