RESUMO
This study aimed to estimate the genetic parameters for somatic cell count (SCC) and the genetic association between SCC and milk production traits using two different methods of SCC normalization. The dataset contained information on 8870 lactation records of 6172 Guzerá dairy cows selected for dual-purpose from 95 herds. The lactation means of SCC were normalized in two ways: (a) SCC1 = log10 (SCC) and (b) SCC2 = log2 (SCC/100) + 3. Multivariate analyses were performed considering milk production traits over the course of 305 days of lactation. Estimates of the variance components and genetic parameters were carried out by the Bayesian inference method, applying Gibbs sampling. Single chains of 2,000,000 iterations were used, with sampling discards of the first 5000 chains and a sampling period of every 50 iterations. The deviation of information criteria (DIC) was used to evaluate the best transformation for standardization of the SCC data, comparing analysis 1 (milk production traits over 305 days and SCC1) with analysis 2 (milk production traits over 305 days and SCC2). According to the data structure of this study, SCC1 normalization was the most efficient method and produced better estimates than normalization by the SCC2 method. The heritability estimates for SCC were low regardless of the transformation method used, indicating a small possibility of expressive genetic gains from the direct selection of these traits. However, the repeatability indicated the potential for increasing heritability estimates if the effects of the permanent environment were reduced. The genetic correlations between the milk yield and SCC traits do not indicate the possibility of a correlated genetic gain from the direct selection of one trait. However, concomitant selection for milk production traits and SCC will likely not affect the individual response either.