Application of Generalized Additive Models to explore minerals in sheep milk

This study explores the complex relationships among ovine milk minerals (Ca, P, Na, K, Mg, K, Cl, respectively) and different factors employing generalized additive mixed models (GAMM). The GAMM included milk yield (MY), parity, and breed as parametric terms, and casein, fat, lactose, pH, SCS, DIM, and sampling day as smooth functions. The objectives were to investigate how these factors could affect minerals in sheep milk and to assess whether their patterns change over time and across different concentrations of major milk components. The GAMM identified distinct patterns in the mineral concentrations between Comisana and Massese breeds, with the Massese ewes having less P, Mg, K, and Cl compared with the Comisana. Moreover, these minerals were also affected by DIM; Mg, S, and Cl were influenced by parity; and P, K, and Na changed across MY levels. Regarding milk components, all the minerals were affected by casein, fat, and lactose concentrations (excluding P for fat). Milk pH was important for Ca, K, Na, and Cl, whereas SCS affected the variability of all minerals except Ca. This study provided valuable insights on the variability of macrominerals in sheep milk, by using GAMM and examining the trajectory of each element across factors as breed, MY, parity, and DIM, as well as across various concentrations of major milk components and their interactions. The dynamic nature of milk mineral content was evident through temporal variability, likely driven by dietary changes, environmental fluctuations, and physiological adaptations, as well as synergistic and antagonistic interactions between milk components and fixed factors. These findings enhance understanding of mineral composition in sheep milk, providing a comprehensive framework for future research on milk quality, animal health, and cheesemaking properties.