Predicting dietary fibre functional properties from their chemical structure: a Raman and FTIR investigation

Dietary fibres display substantial structural heterogeneity which modulate physiological processes across multiple organ systems. Although structure–functional–health relationships are increasingly recognised, quantitative models linking molecular signatures to functional performance remain limited. This study applied Raman and Fourier-transform infrared spectroscopy (FTIR) to acquire vibrational fingerprints from ten dietary fibres, followed by chemometric processing via principal component analysis, linear discriminant analysis, and partial least squares regression. Chemometric analyses robustly differentiated fibre samples according to chemical composition. Regression models constructed from Raman and FTIR spectra established strong structure–function correlations in the calibration data (R 2 > 0.9). However, external validation with five independent fibre samples showed that the models had lower predictive power. Raman spectroscopy delivered accurate predictions of water holding capacity in three samples and water swelling capacity for two, whereas FTIR accurately predicted water swelling capacity of three samples and water holding capacity for one. Oil holding capacity was not predictable from both techniques, showing limitations in capturing lipophilic interactions via vibrational modes alone. These results demonstrate the utility of vibrational spectroscopy coupled with chemometrics as a rapid, non-destructive platform for screening dietary fibre functional potential. Future models can be refined by integrating supplementary properties such as particle size, porosity, or crystallinity. This methodology offers a valuable tool for accelerating the identification of therapeutic fibres in functional food development.