Mid infrared and fluorescence spectroscopies coupled with factorial discriminant analysis technique to identify sheep milk from different feeding systems

in Food Chemistry (2011), 127(2), 743-748

Mid infrared spectroscopy (MIR) combined with multivariate data analysis was used to discriminate between ewes milk samples according to their feeding systems (controls, ewes fed scotch bean and ewes fed ... [more ▼]

Mid infrared spectroscopy (MIR) combined with multivariate data analysis was used to discriminate between ewes milk samples according to their feeding systems (controls, ewes fed scotch bean and ewes fed soybean). The MIR spectra were scanned throughout the first 11 weeks of the lactation stage. When factorial discriminant analysis (FDA) with leave one-out cross-validation was applied, separately, to the three spectral regions in the MIR (i.e. 3000-2800, 1700-1500 and 1500-900 cm(-1)), the classification rate was not satisfactory. Therefore, the first principal component (PCs) scores (corresponding to 3, 10 and 10 for, respectively, the 3000-2800, 1700-1500 and 1500-900 cm(-1)) of the principal component analysis (PCA) extracted from each of the data sets were pooled (concatenated) into a single matrix and analysed by FDA. Correct classification amounting to 71.7% was obtained. Finally, the same procedure was applied to the MIR and fluorescence data sets and 98% of milk samples were found to be correctly classified. Milk samples belonging to control and soybean groups were 100% correctly classified. Regarding milk samples originating from the scotch bean group, only 2 out of 33 samples were misclassified. It was concluded that concatenation of the data sets collected from the two spectroscopic techniques is an efficient tool for authenticating milk samples according to their feeding systems, regardless of the lactation stage. (C) 2011 Published by Elsevier Ltd. [less ▲]

Milk production of imported heifers and Tunisian-born Holstein cows

in American-Eurasian Journal of Agronomy (2009), 1(3), 36-42

Test day (TD) records of milk, fat and protein yields and somatic cell scores (SCS) were studied in Holstein cows in Tunisia. There were 43114, 32923 and 24633 lactation records collected on first, second ... [more ▼]

Test day (TD) records of milk, fat and protein yields and somatic cell scores (SCS) were studied in Holstein cows in Tunisia. There were 43114, 32923 and 24633 lactation records collected on first, second and third parity cows between 1992 and 2004 in 182 herds. Records were of cows born in Tunisia (22000 cows) and those imported from Europe (10830 cows) and North America (850 cows). Variation of total days in milk (DIM) per lactation was studied in function of the herd, calving year x calving season interaction and the origin of the cow. Test-day records were analyzed using a linear model that included calving year x calving season and herd x test-day date interactions, calving season, calving year and origin of the cow. The effective length of lactation was affected by all factors included in the model (p< 0.0001) in all lactations. Test- day milk, protein and fat yields and TD SCS varied (p< 0.01) with management and climatic factors (calving year x calving season and herd x test-day date interactions and year and season of calving). A cow produced 18.8 kg, 0.61 kg and 0.58 kg of milk, fat and protein yields on a daily basis in all lactations, respectively. Average SCS was 2.8 in the three lactations. The origin of the cow was an important (p < 0.05) source of variation for DIM, yields and SCS in all lactations except for first lactation cell scores (p>0.05). Cows born in Tunisia seemed to perform better than imported cows in the first lactation while imported cows showed clearly better performances in later lactations. North American cows produced the highest yields and had the lowest SCS among all cows in the second and third lactations. Imported high producing cows seemed able to adjust to Tunisian management conditions following their first lactation. [less ▲]