Association study between milk traits and mozzarella yield using Bayesian inference

Abstract

The Buffalo milk is rich in solid components and this property is very appreciated by the milk industry to produce 'mozzarella' cheese. A total of 4,834 lactation records were utilized to estimate heritabilities and genetic and phenotypic correlations for 305-d milk yield (MY305), fat (%F) and protein (%P) percentage and, mozzarella chesse production (MoY). The (co)variance components were estimated using multi-trait analysis by Bayesian inference method, through Gibbs sampling. The model included the fixed effects of contemporary groups (herd-year and calving season), number of milking (two levels) and age of cow at calving as covariable (quadratic and linear effect). The additive genetic, permanent environmental and residual effects were included as random effects in the model. The posterior means of heritability distributions for MY305, %F, P% and MoY were 0.25; 0.30; 0.38 and 0.23, respectively. Genetic correlation estimates between MY305, %F and P% were moderate and negative, being the other genetic correlation estimates obtained positive and moderate to high, between %P and MP305 (0.19) and, between MY305 and MP305 (0.95). The milk yield and their components and MP305 have enough genetic variation to show response to selection. The MY305 is an adequate selection criterion to improve MoY.