Investigating Seasonal Variation in Composition and Heat Stability of Bovine Milk from a Pasture Based System using Mid Infrared Spectroscopy and Chemometrics

Ireland's seasonal dairy system results in varied milk composition over the lactation period, affecting milk processing, ingredient functionality and quality of dairy products. This thesis investigates these compositional changes, develops prediction models from spectroscopy data, and uses FTIR and fluorescence spectra as indices for the onset of protein denaturation and destabilisation upon heating. Increases in total protein (3.52 – 3.73 %), casein (2.76 – 2.94 %) and fat (4.22 - 4.48 %) content were observed between 2013 and 2020, while lactose content increased to a maximum of 4.85 % in 2017, then decreased to 4.63 % in 2020. This resulted in an increased protein-to-protein plus lactose ratio from 0.432 in 2013 to 0.443 in 2019, thus the quantity of lactose or milk permeate required for standardisation also increased. The effects of lactation were further examined through the development of prediction models based on the stage of lactation using Irish milk FT-IR spectra acquired. Findings revealed spectral differences in protein and lactose regions between early and late lactation stages. Local prediction models, developed using Partial Least Squares (PLS) and LASSO regression, achieved higher accuracy for total protein and fat content (R² = 0.993 and RMSE = 0.051) compared to commercial models (R² = 0.983 and RMSE = 0.111), highlighting the effectiveness of lactation-specific models in predicting milk composition in Ireland's seasonal dairy context. This thesis also investigated the in-situ molecular changes in milk proteins upon heating, a key factor affecting processing, product quality and shelf life. ATR-FTIR spectroscopy revealed that protein denaturation rates varied with pH, with faster changes observed at pH 6.2 compared to pH 6.8 or pH 7. These differences were determined using a novel approach of measuring time-dependent changes in the spectral regions at 1624 cm-1, related to β-sheet structure and associated with denaturation of β-Lg, and at 1680-1695 cm-1, a region associated with -sheets involved in aggregation. The findings are supported by an increase at 1072 cm⁻¹ associated with colloidal calcium phosphate (CCP), indicating a transition of calcium phosphate from the serum phase to the casein micelle in samples at pH 7. Combining fluorescence and-FTIR spectra, turbidity and particle size data confirmed that changes in protein structure could be correlated with the thermal stability of milk standardised with either lactose or permeate. Skim milk samples containing permeate were less heat stable , decreased in particle size when heated at pH 6.8 compared to those containing lactose, which had an increase in particle size when heated. The thesis shows that a combination of spectrometric techniques can be used to determine heat-induced functionality in a seasonal skim milk supply. The thesis also provides insights into the interrelationships between molecular changes during heating of milk with different composition, having wide application for in-process testing and finished product quality.