Use of Fourier transform infrared (FTIR) spectroscopy to determine the health-promoting index (HPI) of cow's milk

Abstract

Fourier transform infrared (FTIR) spectroscopy steadily has been gaining popularity for the analysis of milk components including total fat, protein, lactose, and ketones. To date, there have been few descriptions of the use of FTIR spectroscopy for the determination of milk fatty acid composition. The first objective of this thesis was to develop a method utilizing FTIR spectroscopy for prediction of the health-promoting index (HPI) of milk. The HPI is known to affect the relative softness of butter (Bobe et al., 2003); so, this model could be used to predict which cows’ milk could be designated for generation of niche products. To accomplish this objective, milk was collected from 281 cows for analysis. One set of samples was esterified into butyl esters for analysis via gas chromatography (GC), and the lipid was extracted from another set to analyze via FTIR spectroscopy. The GC-derived values and FTIR spectra were used to make predictive models by using the partial least squares (PLS) procedure of SAS. The models subsequently were used to generate predictions of milk composition by using FTIR spectra from three additional sets of 135 cows. The predictive model generated for HPI did not generate values for data not included in the predictive model satisfactorily (mean = 0.266, root mean-squared error of prediction (RMSEP) = 0.122); so, predictive models were created for additional milk fatty acid combinations. The predictive models generated estimates for several milk composition parameters relatively well. For example, the model generated to determine the ratio of unsaturated to saturated fatty acids is able to predict with a mean = 0.347 and RMSEP = 0.096. With further development, the use of FTIR