The Feasibility of Nonlinear Models to Describe the Milk Somatic Cell Score of Iranian Holstein Cows throughout Different Lactation Periods



The main objective of the present study was to explore the feasibility of nonlinear mixed models to describe the Somatic Cell Score (SCS) lactation curves and to compare the fit of four nonlinear vs. two linear models when applied to SCS lactation records in Iranian Holstein cows. The Animal Breeding Center of Iran provided the SCC data. The data consisted of 445077 test-day observations from 69124 first to fourth lactation Iranian Holstein cows recorded during the years 2002 to 2007. Six different mathematical functions including Incompelete gama function (Wood), Morant and Gnanasakthy, Ali and Schaeffer function, Wilmink, Rook as well as Nelder functions were fitted to data. The functions were compared based on adjusted R-square and Mean Standard Error (MSE). The results indicated that in the first as well as in the second lactations, Ali and Schaeffer functions described the SCS lactation curve more appropriately than other functions. However in the third and forth lactation, Morant function was better filted than the other five functional forms. Therefore, it can be concluded that the best function is a relative term and depends on the lactation period.