Efficiency of genome-wide association study in open-pollinated populations

• Ano de defesa: 2016
• Autor(a) principal: Mundim, Gabriel Borges
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: eng
• Instituição de defesa: Universidade Federal de Viçosa
• Palavras-chave em Português: Genética vegetal; Genética quantitativa; Milho - Melhoramento genético; Genética Quantitativa
• Link de acesso: http://www.locus.ufv.br/handle/123456789/7933

Most papers about genome-wide association studies (GWAS) with plant species published until now have employed inbred lines panel and almost no information on GWAS in open-pollinated populations was found in literature. Therefore, the objectives of this study were (i) to present theoretical aspects, potential and limitations of GWAS in open-pollinated populations; (ii) to analyze the influence of QTL heritability and population sample size on GWAS with open-pollinated populations; and (iii) to compare the efficacy of GWAS on QTL detection in open-pollinated populations, inbred lines panel and recombinant inbred lines (RILs). Fifty samples of populations with linkage disequilibrium (LD) were simulated, considering sample sizes of 400 and 200 individuals, and 10,000 SNPs, 10 QTL and 90 minor genes were randomly distributed in 10 chromosomes, with an average SNP density of 0.1 cM. The phenotypic values simulated refer to three popcorn traits with different degrees of dominance, considering broad sense heritabilities of 0.4 and 0.8. The scenarios were compared based on power of QTL detection, number of false-positive associations, bias in the estimated QTL position and range of the significant SNPs for the same QTL. Results evidenced that when the LD between a QTL and one or few markers is restricted to SNPs very close or within the QTL, the GWAS in open-pollinated populations can be highly efficient (up to 80% power of QTL detection with reduced number of spurious associations), depending mainly on the population sample size and trait heritability. For inbred lines panel, correcting for population structure, the GWAS achieved the highest power of QTL detection (around 96%), associated with the smallest number of spurious associations and bias. Under low heritability and reduced sample size, GWAS are ineffective for open-pollinated populations, inbred lines panel and RILs.