Incorporation of genomic information in the development of economic indices for dairy cattle selection

The economic efficiency in dairy cattle is related to the use of animals with good performance in production, reproduction, health and longevity. This way, the selection index can be an important tool to increase profitability in this system, since it allows sire selection for multiple traits simultaneously, considering the relationship between them and their economic relevance for the activity. Also, the recent availability of genomic data has permitted to expand the coverage and accuracy of selection indexes by increasing the number and quality of the information considered. In this context, two studies were developed. In the first, the aim was to estimate genetic parameters and breeding values (BV) for milk production and quality traits, including the genomic information in genetic evaluation. Measures of age at first calving (AFC), milk yield (MY), somatic cells score (SCS) and percentages of fat (FP), protein (PP), lactose, casein, and fatty acids in milk of 4,218 cows as well as the genotypes of 755 of these cows for 57,368 single nucleotide polymorphisms (SNPs) were used. The variance components and BV were estimated from a mixed animal model which included the effects of contemporary groups, lactation order, days in lactation, and the additive genetic, permanent environmental and residual effects. Two approaches were developed: a traditional approach, in which the relationship matrix is based on pedigree information; and a genomic approach, in which the matrix is constructed by combining the pedigree and SNP information. The heritabilities ranged from 0.07 to 0.39. Genetic correlations between MY and milk components were between -0.45 and -0.13 whereas high and positive correlations were estimated between FP and fatty acids. The use of the genomic approach did not change genetic parameter estimates; however, there was an increase between 1.5% and 6.8% in BV accuracy; except for AFC, for which a reduction of 1.9% was observed. In the second study, the aim was to incorporate genomic information in the development of economic indexes for sire selection. In this, the BV for MY, FP, PP, total unsaturated fatty acids content (UFA), SCS and herd life were combined in selection indexes weighted by economic values estimated under three payment scenarios: exclusively by milk volume (PAY1); by milk volume and milk components (PAY2); and by milk volume and milk components including UFA (PAY3). These BV were predicted by using phenotypes of 4,293 cows and genotypes of 755 animals in a multi-trait model under traditional and genomic approaches. The use of genomic information influenced the estimates of variance component, BV and response to selection. However, the rank correlations between the approaches were high in all scenarios, with values between 0.91 and 0.99. Differences were mainly observed among PAY1 and the other scenarios, with correlations between 0.67 and 0.88. The relative importance of the traits and the profile of the best animals were sensitive to the scenario considered. Thus, it is essential to consider the economic values of the traits in genetic evaluation and selection decisions. (AU)