Efeito na estrutura populacional e no progresso genético de uma pequena população simulada a partir do uso de diferentes softwares de acasalamentos dirigidos

• Ano de defesa: 2017
• Autor(a) principal: Montenegro, Assis Rubens
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Palavras-chave em Português: Acasalamentos otimizados; Programação linear; Ótima contribuição genética; Algoritmo diferencial evolucionário
• Link de acesso: http://www.repositorio.ufc.br/handle/riufc/47006

Strategies to promote genetic progress or preserve genetic diversity in small populations may change with negative effects of inbreeding. The increase in this parameter occurs due to the BLUP methodology, which tends to favor the selection of animals within the best families, associated with the reproductive techniques that allow the greater dissemination of this genetic material. The increase in the second case is due to the reduced effective population size, which makes the herds more susceptible to genetic drift and inbred mating. In both cases, levels of inbreeding tend to increase over generations, because of the intensity of selection or non-systematic factors, respectively. Mating and selection are the achievable tools to achieve selection goals. Therefore, the search for algorithms that assist the genetic improvement strategy is a recurring concern. Mating optimization approaches such as linear programming (SGRmate software), the optimal genetic contribution, using the Lagrangian multipliers (Gencont software) and, more recently, the evolutionary algorithm (Mate Selection software) have been suggested. In this study, we compared these three methodologies in simulated data that mimicked small- closed populations. In a first scenario (T10) 10 males and 50 females were selected with the highest genetic values in each selection process; in a second scenario (T17), 17 males and 50 females were selected at each optimization. Both scenarios were evaluated over ten generations. It was also evaluated a random mating scenario (ACASO) as a comparison parameter. Algorithms optimized the objective function in order to achieve the greatest genetic progress for an inbreeding limit of 10%, selecting the necessary number of males and forming the reproductive pairs, except for Gencont, whose objective function was only to minimize the coancestry. All softwares generated populations with similar genetic progress. Regarding the population structure, Mate Selection generated populations with the highest levels of inbreeding, similar to the ACASO scenario. By contrast, this was the software that best controlled mating between relatives. Gencont produced populations with intermediate levels of inbreeding. Finally, SGRmate software was the one that maintained the lowest levels of inbreeding due to the greater number of males selected and equal proportionality of combined use with the formation of reproductive pairs. We concluded that the use of linear programming implemented in SGRmate software, was more efficient in maintaining the genetic diversity of small- closed populations.