Crews D H, Franke D E
Department of Animal Science, Louisiana Agricultural Experiment Station, Louisiana State University Agricultural Center, Baton Rouge 70803, USA.
J Anim Sci. 1998 Jul;76(7):1803-9. doi: 10.2527/1998.7671803x.
Heterogeneity of carcass trait variances due to level of Brahman inheritance was investigated using records from straightbred and crossbred steers produced from 1970 to 1988 (n = 1,530). Angus, Brahman, Charolais, and Hereford sires were mated to straightbred and crossbred cows to produce straightbred, F1, back-cross, three-breed cross, and two-, three-, and four-breed rotational crossbred steers in four non-overlapping generations. At weaning (mean age = 220 d), steers were randomly assigned within breed group directly to the feedlot for 200 d, or to a backgrounding and stocker phase before feeding. Stocker steers were fed from 70 to 100 d in generations 1 and 2 and from 60 to 120 d in generations 3 and 4. Carcass traits included hot carcass weight, subcutaneous fat thickness and longissimus muscle area at the 12-13th rib interface, carcass weight-adjusted longissimus muscle area, USDA yield grade, estimated total lean yield, marbling score, and Warner-Bratzler shear force. Steers were classified as either high Brahman (50 to 100% Brahman), moderate Brahman (25 to 49% Brahman), or low Brahman (0 to 24% Brahman) inheritance. Two types of animal models were fit with regard to level of Brahman inheritance. One model assumed similar variances between pairs of Brahman inheritance groups, and the second model assumed different variances between pairs of Brahman inheritance groups. Fixed sources of variation in both models included direct and maternal additive and nonadditive breed effects, year of birth, and slaughter age. Variances were estimated using derivative free REML procedures. Likelihood ratio tests were used to compare models. The model accounting for heterogeneous variances had a greater likelihood (P < .001) than the model assuming homogeneous variances for hot carcass weight, longissimus muscle area, weight-adjusted longissimus muscle area, total lean yield, and Warner-Bratzler shear force, indicating improved fit with percentage Brahman inheritance considered as a source of heterogeneity of variance. Genetic covariances estimated from the model accounting for heterogeneous variances resulted in genetic correlations of or near unity. These results suggest that different genetic values be considered for genetic evaluation of carcass yield and shear force traits from steers with different degrees of Brahman inheritance.
利用1970年至1988年期间纯种和杂交阉牛的记录（n = 1530），研究了婆罗门牛遗传水平对胴体性状方差的异质性。安格斯、婆罗门、夏洛来和赫里福德公牛与纯种和杂交母牛交配，在四个不重叠的世代中产生纯种、F1代、回交、三品种杂交以及二品种、三品种和四品种轮回杂交阉牛。断奶时（平均年龄 = 220天），阉牛在品种组内被随机分配，直接进入饲养场饲养200天，或者在育肥前先经过一个育成和架子牛阶段。第1代和第2代的架子牛饲养70至100天，第3代和第4代饲养60至120天。胴体性状包括热胴体重、第12 - 13肋处皮下脂肪厚度和背最长肌面积、胴体重校正背最长肌面积、美国农业部产量等级、估计总瘦肉产量、大理石花纹评分以及沃纳 -布拉茨勒剪切力。阉牛根据婆罗门牛遗传水平分为高婆罗门牛（50%至100%婆罗门牛血统）、中婆罗门牛（25%至49%婆罗门牛血统）或低婆罗门牛（0%至24%婆罗门牛血统）。针对婆罗门牛遗传水平拟合了两种动物模型。一种模型假设婆罗门牛遗传组对之间的方差相似，第二种模型假设婆罗门牛遗传组对之间的方差不同。两种模型中的固定变异来源包括直接和母体加性及非加性品种效应、出生年份和屠宰年龄。使用无导数REML程序估计方差。用似然比检验比较模型。对于热胴体重、背最长肌面积、体重校正背最长肌面积、总瘦肉产量和沃纳 - 布拉茨勒剪切力，考虑方差异质性的模型比假设方差齐性的模型具有更高的似然性（P < 0.001），这表明将婆罗门牛遗传百分比视为方差异质性来源时拟合效果更好。从考虑方差异质性的模型估计的遗传协方差导致遗传相关系数为1或接近1。这些结果表明，对于不同婆罗门牛遗传程度的阉牛，在胴体产量和剪切力性状的遗传评估中应考虑不同的遗传值。
