Estudo genômico da produção de leite e seus constituintes em bovinos da raça Guzerá

• Ano de defesa: 2015
• Autor(a) principal: Santos, Daniel Jordan de Abreu [UNESP]
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Estadual Paulista (Unesp)
• Palavras-chave em Português: Bovino; Genômica; Guzera (Zebu); Leite - Produção; Genetica animal; Genomics
• Link de acesso: http://hdl.handle.net/11449/136113; http://www.athena.biblioteca.unesp.br/exlibris/bd/cathedra/15-02-2016/000858092.pdf

Commercial panels containing thousands of SNPs revolutionized genetic studies in livestock, especially with the genome-wide association study and genomic selection. The genomic selection is also directly applied to breeding programs enabling increasing on accuracy of the genetic evaluation for quantitative traits such as milk yield and its constituents. Thus, the aim of this thesis was to study the distribution of frequencies of polymorphisms and the linkage disequilibrium (LD) in the Guzerá cattle; was also to study the association of the markers with the dairy traits and to compare different models for genomic evaluation with different prior distributions for the effect of markers. The LD and the effective size were estimated using a 50 k Illumina® panel for 50 sires and 853 cows. For 100 kb the average of r2 was 0.16. The density of this 50 k panel was considered sufficient to provide LD between chromosomal segments for estimation of genomic breeding values. The effective size was reduced over the course of generations, indicating increase intensity of selection. The low estimate for effective size for recent generations indicated the importance of considering inbreeding for mating to maintain the genetic diversity of this breed. Genomic evaluations were performed by GBLUP, BayesC, BayesCπ, Lasso and two multi-trait models for milk yield (MY), fat (MF) and protein (MP). The accuracies for predictions ranged from 0.65 to 0.80. Bayesian prediction, GBLUP and multi-traits models were equivalent for predict the genomic breeding values. However, the best option was GBLUP considering overall fit, lower computational requirement and facility of convergence. The results indicated that genomic predictions were adequate to assist animal selection. In other study, the percentages of explained variance for each SNP were calculated. Windows, composed of seven adjacent SNPs deviated substantially from the other genomic regions were ...