Effects of milk fat,casein, and serum protein concentrations on sensory properties of milk-based beverages

Our goal was to determine the effect of systematically controlled variation in milk fat, true protein, casein, and serum protein concentrations on the sensory color, flavor and texture properties, instrumental color and viscosity, and milk fat globule size distribution of milk-based beverages. Beverage formulations were based on a complete balanced 3-factor (fat, true protein, and casein as a percentage of true protein) design with 3 fat levels (0.2, 1.0, and 2.0%), 4 true protein (TP) levels (3.00, 3.67, 4.34, and 5.00%) within each fat level, and 5 casein as a percentage of true protein (CN%TP) levels (5, 25, 50, 75, and 80%) within each protein level (for a total of 60 formulations within each of 2 replicates). Instrumental measures of Hunter L and a values and Commission Internationale de l'Éclairage (CIE) b* values, instrumental viscosity, particle size, flavor, sensory texture and sensory appearance evaluations were done on each pasteurized/homogenized beverage formulation. Within each of the 3 fat levels, higher serum protein concentration drove higher aroma intensity, sweet aromatic, cooked/sulfur, cardboard/doughy flavors, and sensory yellowness scores, whereas higher casein concentration drove higher instrumental viscosity in milk protein beverages. Increasing serum protein concentration increased yellowness, sweet aromatic, aroma intensity, cooked/sulfur, and cardboard/doughy flavors across all fat levels and also had the largest effect on L, a, and b* values, sensory whiteness, and opacity within each fat level. Increases in true protein increased throat cling and astringency intensities. Increases in fat concentration were correlated with higher L, a, and b* values, larger particle size, and increased sensory whiteness, mouth coating, cooked/milky, and milkfat flavors. Multiple linear regression of L, a, and b* values produced better predictions of sensory whiteness and yellowness of pasteurized milk protein beverages than simple linear regression of L or b* values, respectively. Formulating milk protein beverages to a higher true protein level increased astringency regardless of fat level. When formulating milk protein beverages, a product developer has a wide range of milk-based protein ingredient choices that differ in price and change price relationship across time. Understanding the expected relative effect of different milk protein ingredients on the textural and flavor characteristics of milk-based beverages could be used to help guide product reformulation decisions and ingredient choices to achieve a specific sensory profile while controlling total beverage ingredient cost.