Estimation of Casein from Total Protein in Commingled Milk

Abstract

Casein variation as a percentage of total protein and true protein in 146 commingled milk samples was investigated. A micro-Kjeldahl method was used to measure total protein (total nitrogen × 6.38), true protein [(total nitrogen − nonprotein nitrogen) × 6.38], and casein [(total nitrogen − noncasein nitrogen) × 6.38]. Samples were collected at seven Vermont cheese plants for 11 mo. Percentage casein of total protein and of true protein differed significantly among plants and months with no interactions of plant by month. Casein varied considerably less as a percentage of true protein than of total protein. Although highly significant, differences among plants and months were small. Among the 146 samples percents ranged from 74.9 to 79.3 casein of total protein and 80.4 to 83.7 casein of true protein. Four regression models for estimating casein from total protein were evaluated. Total protein accounted for 87.8% of variation of casein, whereas true protein accounted for 93.1%. Two factors - regression of casein on total protein and months explained 94.3% of variation of casein, with a standard error of estimate of .018% casein. In this model, explained variability increased by only 1% when cheese plant was included as an independent variable. © 1983, American Dairy Science Association. All rights reserved.