Is it possible to predict milk processing into butter using infrared spectroscopy?

Description of the subject. Given the current low price of milk, a lot of producers have decided to process their milk into products with a higher added-value, including butter. However, all milks are not suitable to be transformed into butter. It would thus be useful to be able to predict milk processing properties. Objectives. The aim of this paper was to study the ability of milk to be processed into butter using infrared spectrophotometry. Method. A normalized protocol for the production of butter was developed. Milk samples (n = 110) collected between 2013 and 2016 were analyzed by near and medium infrared spectrometry (315 spectra). Butter samples were also analyzed by visible-near infrared spectrometry (220 spectra). Composition of the products was subsequently assessed using validated prediction equations. Principal components analyses were performed to discriminate samples. Results. Butter properties seemed to be influenced by seasons and feedings. Water content and color parameters could be predicted on the basis of butter infrared spectra. Conclusions. It was possible to correlate butter characteristics with milk properties. However, it was not possible to predict butter characteristics on the basis of milk near infrared spectra. It could be interesting to try predictions from milk medium infrared spectra.