Development of a near Infrared Method for the Quantification of the Main Classes of Fatty Acids Obtained from Raw Milk by Solvent-Free Extraction. Issue 5 (October 2013)

• Record Type: Journal Article
• Authors: Cabassi, Giovanni; Povolo, Milena; Pelizzola, Valeria; Monti, Lucia; Genorini, Emiliano; Contarini, Giovanna
• Is Part Of: Journal of near infrared spectroscopy. Volume 21:Issue 5(2013)
• Page Start: 395
• Page End: 408
• Publication Date: 2013-10
• Subjects: milk fatty acid profile -- fat extraction -- saturated fatty acids -- monounsaturated fatty acids -- polyunsaturated fatty acids -- trans fatty acids; Near infrared spectroscopy -- Periodicals; 543.5
• DOI: 10.1255/jnirs.1074
• Languages: English
• ISSNs: 0967-0335

Abstract

The determination of the fatty acid (FA) profile of milk fat, which is the most complex dietary fat, is more and more for both the definition of nutritional labelling and the study of the genetic variability in bovine milk fat quality. For these reasons, the development of infrared techniques for the rapid evaluation of FA profiles is gaining much wider interest. In this research, a method for the rapid determination of saturated, unsaturated, monounsaturated, polyunsaturated and trans FAs, expressed on 100g of total milk FAs, is reported. The FA profile was estimated from the transmission spectrum (optical pathlength of 6 mm) obtained from the fat extracted by hot centrifugation, without solvents. Authentic triglyceride standards were used in order to perform the spectral band assignments. The whole data set included 135 raw milk samples collected from cows fed either a total mixed ration or grazed on mountain pastures. Ninety five samples were selected for model development (calibration set); the prediction models were then validated by an independent validation set composed by 40 samples. The prediction model for the saturated FA content showed a correlation coefficient of determination ( r 2 ) = 0.96, standard error of prediction ( SEP ) = 0.97 (g 100 g −1 of total FAs), ratio of errors ( RER ) = 22.8 and ratio of performance to deviation ( RPD ) = 4.6. Monounsaturated, polyunsaturated and trans FAs provided r 2 = 0.95, 0.90 and 0.97, RER = 20.4, 12.5 and 32.1 and RPD = 4.2, 2.8 and 5.2, respectively. The ability to predict the concentration of some saturated single FAs, also grouped according to their chain length, was also investigated. Satisfactory screening performances were found for C14, C16 and C18 FAs, as well as for the amount of the groups including medium (12–17 carbon atoms) and long (18–22 carbon atoms) FAs.