Abstract
Near-infrared (NIR) spectroscopy with chemometrics was used to evaluate semi-refined flaxseed protein extract (SRFPE) and predict its protein content. The effect of extraction medium (water vs. alkaline) employed after pulsed electric field (PEF) treatment on protein yield was also studied. Spectral data were validated for nitrogen content against the Dumas and Kjeldahl methods. The root mean square error of prediction (RMSEP) and the ratio of performance deviation (RPD) for optimally preprocessed data were 0.09 % and 8.26 respectively. The spectral shifts suggested protein conformational changes influenced by extraction method, pH and the applied PEF energy input. These results showed the ability of NIR (908–1676 nm) for rapid, accurate and non-destructive protein quantification across diverse sample matrices while demonstrating its potential for streamlining analytical processes.
• MicroNIR spectroscopy with chemometrics enabled rapid protein content prediction.
• Second derivative + standard normal variate preprocessing optimized models.
• Protein content predictions exhibited highest accuracy using the Dumas method.
• Protein conformations differ between alkali, water, and PEF treatments.