Mapeamento de marcas moleculares e identificação de QTL com base em desequilíbrio de fase gamética

• Autor(a) principal: Silva, Admilson da Costa e
• Data de Publicação: 2008
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: LOCUS Repositório Institucional da UFV
• Texto Completo: http://locus.ufv.br/handle/123456789/4665
• topic: Desequilíbrio de ligação gênica; Marcador co-dominante; Simulação; Gene linkage disequilibrium; Co-dominant marker; Simulation; CNPQ::CIENCIAS BIOLOGICAS::GENETICA::GENETICA QUANTITATIVA

In the study of genetic linkage between two marker loci or between markers and a QTL it is usually necessary to have appropriate designed populations (experimental populations). However, for some species such designed populations could be hard to obtain. In this scenario, the linkage disequilibrium (LD) has been used to build genetic map and to identify QTL. In this work, we used simulation to evaluate the efficiency of LD to construct genetic maps (LD-genetic map) of molecular markers as well as to map QTL (LD-QTL mapping) in non- structured families of outbreed populations. As for the LD- genetic map study, we designed genomes with high saturation level (markers apart from each other less than 5 cM), and random saturation level (markers apart from each other ranging from 0 to 20 cM), five linkage groups and 20 markers per linkage group. We simulated compounds of second generation with 200 and 1000 individuals, derived from the mating of two parental populations of size 200, under four levels (1.0, 0.9, 0.8 and 0.7) of difference in the frequency of each marker (marker-frequency difference) between parental populations. These compounds were used to obtain the sampled genomes, which were compared with the designed genomes to measure the efficiency of LD to construct genetic map. In the LD-QTL mapping study, we adopted, with slightly modifications, the steps to obtain the designed genome and the compounds as described above. We designed genomes with ten linkage groups, each with 20 markers, but for linkage group 1 which had just five markers. We distributed 200 genes controlling quantitative traits along the ten linkage groups. However, only the linkage group 1 had a major QTL along with 19 minor genes. All other linkage groups had 20 minor genes each. In the LD-QTL mapping study, for the detection, estimation of substitution effects and of dominance, we idealized three quantitative traits, each being affected by a major QTL. The trait yield (g/ear) had a QTL with positive dominance; the trait expansion capacity (mL/g) had a QTL with bidirectional dominance; and the trait relative liquid growth (%) had a QTL with negative dominance. We used the method of single marker analysis base on regression to model the QTL effects. The analyses of power for QTL identification and substitution effect estimation were done by linear regression. On the other hand, the analysis of dominance effect was done by polynomial regression with the quadratic term included in the model. In the LD-genetic map study the linkage groups were satisfactory recovered, with only few exceptions. The bias in recovering the designed genome increased with the decreasing of the marker- frequency difference between parental populations from 1.0 to 0.7. Based on the percentage of correct ordered markers in the genome, we found that the LD-genetic map was efficient, with correct ordering of markers superior to 86%. Our results support that the most efficient population for LD-genetic map was the one equivalent to an F2 intercross population. With respect to the QTL detection, in general, the power of detection was satisfactory, independent of the quantitative trait simulated. The results of effect substitution of marker revealed that markers closer to the QTL had bigger effects. The detections of dominance effects were efficient only on the compounds from the mating of parental populations with marker-frequency difference superior than 0.9 and sample size of 1.000. The results of dominance direction analyses on the markers for the trait yield were efficient in determining the direction of dominance of QTL, with few exceptions where complete dominance was present. For the trait expansion capacity, the specification of which markers had positive or negative effect was not possible because of the bidirectional nature of the simulated dominance effects. For the trait liquid growth the estimated effects of dominance at markers often revealed the correct dominance effects direction of the QTL. It is important to point out that the results of our LD-QTL mapping study were based on traits in which the phenotypic variance explained by the QTL is between 10 and 30% of the total phenotypic variance.