Estimates of genetic parameter for litter size at different parities in crossbred large white and landrace sows

Abstract

The objectives of this study were to estimate genetic parameters of number of born alive (NBA) in different parities and to examine the appropriateness of the 3 models. The genetic parameters in all models were evaluated. In model I, single-trait animal model, each parity was considered as a different trait. In model II, repeatability model, each trait of different paritys was measured repeatedly. In model I, multiple-trait model, each parity was treated as a different trait. The 6,375 records of NBA from sows that born between 2546 to 2552 were analysed. They were grew by private farms in the central part of Thailand. (Co)variance components and breeding value were estimated by Restricted Maximum Likelihood (REML) and Best Linear Unbiased Prediction (BLUP) method, respectively. The results revealed that heritabilities of NBA from Model I, II, and III were less which ranged from 0.01-0.05. Genetic correlation (rg) between parity were less which ranged from 0.03-0.66.Model comparison by -2LogL values indicate that Model I is the most appropriate model. For the accuracy of estimated breeding values (EBV), Model II had the highest accuracy. The top 25 dams of all three models that were analyzed by the rank correlation between EBV have different ranking. Therefore, Model II was selected as an appropriate model to estimate the population of breeding values pigs for maximizing the response in this study.