This study compares growth traits (weaning weight, slaughter weight and average daily gain) in four maternal lines of rabbits, A, V, H y LP, founded under different criteria, but all of them selected for litter size at weaning at the nucleus of the Polytechnic University of Valencia, until generations 41th, 37th, 10th (only until May 2004) and 6th respectively . A Bayesian approach was used for inference. Two types of comparisons were performed. One was carried out at the origin of the lines, using the complete data set (data from June 1980 to November 2010) and a two-trait analysis. Each growth trait was analysed together with litter size at weaning to avoid bias due to selection. The model used for the growth traits was a mixed model that includes as random effects, the additive, the maternal and the litter of birth. The consideration of the complete data set and the inclusion of the additive effects allow the comparison between the lines at their origin and allowed to predict the differences for given periods (expected differences). The other type of comparisons, at fixed periods of time, was done from March 1997 to August 1998 for comparisons between the lines A, V and H (period 1), and from December 2009 to November 2010 for comparisons between the lines A, V and LP (period 2). Each period of comparison corresponded to the last period when all the lines of each set were housed together at the nucleus, having the same management. These analyses used only the data
corresponding to each period and one-trait models for the growth traits. These models were the same than the ones referred for the comparisons at the origin, but the additive effects were excluded. These comparisons allow estimating the actual differences between the lines at the given periods (observed differences) and comparing them to the expected differences for the same periods. A similarity between observed and predicted differences can be interpreted as an indication of the suitability of the complete model to explain the traits.
The estimates of the genetic correlations between litter size at weaning and growth traits were positive but low. At the origin, significant differences were observed for growth traits between the different lines. The contrast between H and LP lines was not done because they only shared two years-season. The results showed that at the origin, the H and LP lines were the heaviest. These differences may be due to different criteria and processes used to establish these lines. In period 1, the results showed that the H line was the heaviest. In period 2, the results showed that the LP line was the heaviest. In both periods, the observed (computed with the records of each period) and the expected differences (computed with the complete model and data set) were very similar for all traits in the two periods, indicating the suitability of the used models. At each period, the differences between lines for growth traits were smaller than at the origin. This result could be due to a correlated response to the selection for litter size at weaning and to the direct response to a concomitant, non-programmed selection for the growth traits themselves different in intensity between the lines.
|
