Hammami H, Rekik B, Bastin C, Soyeurt H, Bormann J, Stoll J, Gengler N
Gembloux Agricultural University, Gembloux, Belgium.
J Dairy Sci. 2009 Sep;92(9):4604-12. doi: 10.3168/jds.2008-1513.
Milk production data of Luxembourg and Tunisian Holstein cows were analyzed using herd management (HM) level. Herds in each country were clustered into high, medium, and low HM levels based on solutions of herd-test-date and herd-year of calving effects from national evaluations. Data from both populations included 730,810 test-day (TD) milk yield records from 87,734 first-lactation cows. A multi-trait, random regression TD model was used to estimate (co)variance components for milk yield within and across country HM levels. Additive genetic and permanent environmental variances of TD milk yields varied with management level in Tunisia and Luxembourg. Additive variances were smaller across HM levels in Tunisia than in Luxembourg, whereas permanent environmental variances were larger in Tunisian HM levels. Highest heritability estimates of 305-d milk yield (0.41 and 0.21) were found in high HM levels, whereas lowest estimates (0.31 and 0.12, respectively) were associated with low HM levels in both countries. Genetic correlations among Luxembourg HM levels were >0.96, whereas those among Tunisian HM levels were below 0.80. Respective rank orders of sires ranged from 0.73 to 0.83 across Luxembourg environments and from 0.33 to 0.42 across Tunisian HM levels indicating high re-ranking of sires in Tunisia and only a scaling effect in Luxembourg. Across-country environment analysis showed that estimates of genetic variance in the high, medium, and low classes of Tunisian environments were 45, 69, and 81% lower, respectively, than the estimate found in the high Luxembourg HM level. Genetic correlations among 305-d milk yields in Tunisian and Luxembourg HM environments ranged from 0.39 to 0.79. The largest estimated genetic correlation was found between the medium Luxembourg and high Tunisian HM levels. Rank correlations for common sires' estimated breeding values among HM environments were low and ranged from 0.19 to 0.39, implying the existence of genotype by environment interaction. These results indicate that daughters of superior sires in Luxembourg have their genetic expression for milk production limited under Tunisian environments. Milk production of cows in the medium and low Luxembourg environments were good predictors of that of their paternal half-sisters in the high Tunisian HM level. Breeding decisions in low-input Tunisian environment should utilize semen from sires with daughters in similar production environments rather than semen of bulls proven in higher management levels.
利用畜群管理(HM)水平对卢森堡和突尼斯荷斯坦奶牛的产奶数据进行了分析。根据国家评估中畜群检测日期和产犊年份效应的解决方案,将每个国家的畜群分为高、中、低三个HM水平。来自两个群体的数据包括87,734头头胎奶牛的730,810个测定日(TD)产奶量记录。使用多性状随机回归TD模型来估计不同国家HM水平内和跨HM水平的产奶量(协)方差分量。突尼斯和卢森堡的TD产奶量的加性遗传方差和永久环境方差随管理水平而变化。突尼斯不同HM水平间的加性方差比卢森堡小,而突尼斯HM水平下的永久环境方差更大。在高HM水平下,305天产奶量的最高遗传力估计值(分别为0.41和0.21)被发现,而在两个国家的低HM水平下遗传力估计值最低(分别为0.31和0.12)。卢森堡不同HM水平间的遗传相关性大于0.96,而突尼斯不同HM水平间的遗传相关性低于0.80。在卢森堡不同环境下,公牛的排名顺序分别为0.73至0.83,在突尼斯不同HM水平下为0.33至0.42,这表明突尼斯公牛的排名变化很大,而在卢森堡只是有一个尺度效应。跨国环境分析表明,突尼斯高、中、低环境类别中遗传方差的估计值分别比卢森堡高HM水平下的估计值低45%、69%和81%。突尼斯和卢森堡HM环境下305天产奶量之间的遗传相关性在0.39至0.79之间。最大的估计遗传相关性出现在卢森堡中等HM水平和突尼斯高HM水平之间。HM环境中共同公牛估计育种值的秩相关较低,范围为0.19至0.39,这意味着存在基因型与环境的相互作用。这些结果表明,在突尼斯环境下,卢森堡优秀公牛的女儿产奶的遗传表达受到限制。卢森堡中低环境下奶牛的产奶量是其在突尼斯高HM水平下父系半姐妹产奶量的良好预测指标。在突尼斯低投入环境下的育种决策应使用来自在类似生产环境中有女儿的公牛的精液,而不是在更高管理水平下被证明的公牛的精液。