CORRELAÇÃO CANÔNICA DE CARACTERÍSTICAS DE PRODUÇÃO EM VACAS DA RAÇA GIROLANDO

The aim of this work was to estimate the relationship between production and milk quality traits, for to determine which are the most relevant in the selection of dairy cattle through canonical correlation analysis. For this study, a database was used containing milk production records of cows of breed Girolanda in different blood grades, raised to pasture, whose traits were divided into two groups: production (milk production, number of calving, age at first calving) and milk quality (percentage of protein, percentage of fat, kilogram of protein produced and kilo of fat produced). The canonical correlation analysis was used, using the CANCORR procedure of the statistical program SAS UNIVERSITY (2018). To verify the significance of canonical correlation, the maximum likelihood ratio test was used. Results indicate that the sets of production traits and milk quality are dependent. Among the evaluated variables, the lactation and the kilo of protein produced were the most important. Thus, when making selection for higher yields and milk quality, one can use only these two traits. The genetic gain obtained in these two traits will also provide genetic gain in the other traits.