Genome-wide association study of reproduction traits in Nelore cattle, including additional phenotypic information from non-genotyped animals

• Ano de defesa: 2015
• Autor(a) principal: Melo, Thaise Pinto de [UNESP]
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: eng
• Instituição de defesa: Universidade Estadual Paulista (Unesp)
• Palavras-chave em Português: Bovino de corte; Nelore (Zebu); Genômica; Idade; Genetica animal; Métodos estatísticos; Genomics
• Link de acesso: http://hdl.handle.net/11449/123734; http://www.athena.biblioteca.unesp.br/exlibris/bd/cathedra/08-06-2015/000834416.pdf

This dissertation was divided in three chapters, the first one is a literature review about the subject that will be discussed in subsequent chapters. In the second chapter, a genome wide association study (GWAS) for age at first calving (AFC) in Nelore cattle was performed, using real and simulated data, aiming to 1) assess if additional phenotypic information from non-genotyped animals affect QTL mapping of AFC; 2) evaluate, by simulation, if this additional phenotypic information contributes to detect QTLs more precisely for a low heritable complex trait, and with few available genotypes. The third chapter presents a GWAS for heifer rebreeding (HR) in Nelore cattle. In chapter two, GWA studies were performed using Bayes C and weighted single step GBLUP (WssGBLUP) methods and the top 10 marker windows (1Mb) that explained the larger proportion of variance for AFC were identified and further explored. Two scenarios were investigated, one including all females with available phenotypic information (SI scenario, with 43,482 females), and the other including just the females with available genotype (SII scenario, with 1,813 females). Three iterations were performed in WssGBLUP, recomputing the animals and SNPs effect in each subsequent iteration. It was simulated a population mimicking the parameters and the structure of the real dataset. Two different disequilibrium linkage levels (low and high) between adjacent markers were simulated. In chapter three, the data consisted of 142,878 HR phenotypic records and 2,923 genotypes. The GWAS was performed with WssGBLUP method using three different weightings (iterations) for the SNP effects. Total genetic variances were calculated for the top 10 1Mb SNP-windows, detected by each iteration. On each subsequent iteration, the genetic variance was distributed for a smaller number of SNPs, and the SNP effects were recomputed. Genes possibly associated with HR were searched to reinforce the suggestive ...