Multiple regression models for estimating the Zn content in cowʹs milk

Relevance. Regular veterinary and sanitary control of the safety and quality of animal products does not imply zinc testing. Although the role and importance of this element is largely due to its quantity. At the same time, the zinc content in milk is not constant and is due to its content in soils and feeds, and the physiology of a dairy cow. After extensive studies of zinc in milk, we calculated a number of multiple regression equations to predict zinc levels based on its biochemical analysis in order to minimize costs. Methods. The analysis of biochemical parameters of cow’s milk was performed using the MilkoScan 7 / Fossomatic 7 DC system (Denmark). Zinc was studied using an atomic absorption spectrometer with deuterium and Zeeman correction ZEEnit 650 P. Results. The average zinc content in milk was set at 3017.7 mcg/l. The degree of influence of the biochemical analysis data on the resulting variable (Zn) showed the high importance of the variable’s fat mass fraction, freezing point and pH (p = 0.006, 0.0001, 0.00003, respectively). The equation is characterized by a high multiple correlation coefficient (0.92) and is significant according to the F-criterion = 5,41E43, the adjusted value of R 2 = 0.83, which can be considered a good result. Working with regression forecasting models allows for a preliminary assessment of the zinc level in milk according to its biochemical analysis, without additional financial burden on production and better control of its content in milk.