The lactation curves can be modeled by different mathematical functions (Wood, Brody, Wilmink), which present a reduced number of parameters with biological interpretations, such as the initial yield (parameter a), the ascending phase of the curve (parameter b) and the descending phase of the curve (parameter c). Thus, desirable changes in the shape of the curves can be obtained by considering the estimates of these parameters as phenotypic observations in statistical models of genetic evaluation. Currently, given the importance of genomic information for Animal Breeding, the use of these phenotypes can be expanded to more sophisticated models, such as the Genome Wide Association Study (GWAS). Therefore, this objective project to estimate lactation curves of buffaloes for the identification of earlier and more efficient animals for selection purposes; identify and select the best animals via genomic genetic values for the parameters of the lactation curves; comparing nonlinear models for genomic evaluation of lactation curves; in addition to identifying chromosomal regions and candidate genes related to the parameters of the lactation curves by GWAS. The data come from approximately 12 farms participating in the dairy control program of the Department of Animal Science of the Universidade Estadual Paulista, Campus Jaboticabal, São Paulo, located in the states of Ceará, Rio Grande do Norte and São Paulo, approximately 3,943 animals born between 1971 and 2014 will be used. In addition, 706 animals were genotyped using a panel of 90K Axiom® Buffalo Genotyping (Affymetrix). The phenotypic observations of five characteristics will be considered, being: milk production on the day of control, accumulated milk production, fat percentage, protein percentage and somatic cell count. The Bayesian hierarchical model will be applied first, to analyze the individual lactation curves based on the nonlinear regression models, after the parameters of the nonlinear models will be modeled by a linear multitrait model. The models will be compared through their predictive ability through cross validation analysis. Once the genomic values of each animal are obtained for the model parameter estimates and the best lactation model chosen, it will be possible to obtain the "genomic lactation curves" for each animal.
News published in Agência FAPESP Newsletter about the scholarship: