Using multivariate analysis to explore the relationships between color, composition, hygienic quality, and coagulation of milk from Manchega sheep

This study explores the relationships between composition, hygienic quality, and color values of milk to evaluate whether colorimetry could be used as a valid predictor of the quality of raw milk and its coagulation. For this purpose, we performed analyses in 1,200 individual samples of Manchega sheep milk from 4 flocks from the region of Castilla-La Mancha, Spain. Based on the measured variables, we determined the differences and similarities between coagulating and noncoagulating samples using discriminant analysis techniques. The variables with a higher discriminant ability were lactose content, somatic cell score, pH, and the color values lightness (L*) and red/green value (a*). The model based on color values showed a predictive ability similar to that found in the model based on milk composition and hygienic quality. Canonical correlation analysis allowed us to explore the relationships between both sets of variables. Canonical correlations for the first and second pair of canonical values were 0.794 and 0.438, respectively. Both values were significant and represented 92.82% of the observed variability. The correlation structure showed that color values had a strong correlation with fat and protein content and with total solids, and they had a weak correlation with lactose content and somatic cell score. The 2 first combinations of standardized canonical variability could be considered a predictable measure of the composition and, to a lesser extent, the hygiene of milk. Measurement of color values could be a rapid and effective means of supplementing standard analyses when determining the coagulation ability of Manchega sheep milk.