Dairy records from the Dairy Recording Service of Kenya were classified into low, medium and high production systems based on mean 305-day milk yield using the K-means clustering method. Milk and fertility records were then analysed to develop genetic evaluation systems accounting for genotype-by-environment interaction between the production systems. Data comprised 26,638 lactation yield, 3,505 fat yield, 9,235 age at first calving and 17,870 calving interval records from 12,631 cows which were descendants of 2,554 sires and 8,433 dams. An animal model was used to estimate variance components, genetic correlations and breeding values for the production systems. Variance components increased with production means, apart from genetic group variances, which decreased from the low to the high production system. Moderate heritabilities were estimated for milk traits (0.21-0.27) and fat traits (0.11-0.38). Low heritabilities were estimated for lactation length (0.04-0.10) and calving interval (0.03-0.06). Moderate heritabilities (0.25-0.26) were estimated for age at first calving, except under the high production system (0.05). Within production systems, lactation milk yield, 305-day milk yield and lactation length had high positive genetic correlations (0.52-0.96), while lactation milk yield and lactation length with age at first calving had negative genetic correlations. Milk yield and calving interval were positively correlated except under the low production system. The genetic correlations for lactation milk yield and 305-day milk yield between low and medium (0.48 ± 0.20 and 0.46 ± 0.21) and low and high production systems' (0.74 ± 0.15 and 0.62 ± 0.17) were significantly lower than one. Milk yield in the low production system is, therefore, a genetically different trait. The low genetic correlations between the three production systems for most milk production and fertility traits suggested that sires should be selected based on progeny performance in the targeted production system.

中文翻译：

---

估算肯尼亚低、中、高乳制品生产系统内和之间的牛奶和生育性状遗传参数，以解释基因型与环境的相互作用

使用 K 均值聚类方法，根据平均 305 天的产奶量，将来自肯尼亚乳制品记录服务的乳制品记录分为低、中和高生产系统。然后分析牛奶和生育记录，以开发遗传评估系统，以解释生产系统之间基因型与环境的相互作用。数据包括来自 2,554 头公牛和 8,433 头母牛的后代的 12,631 头母牛的 26,638 条泌乳产量、3,505 条脂肪产量、9,235 头产犊年龄和 17,870 条产犊间隔记录。动物模型用于估计生产系统的方差分量、遗传相关性和育种值。除了遗传组方差从低生产系统到高生产系统减少外，方差分量随着生产手段的增加而增加。估计牛奶性状 (0.21-0.27) 和脂肪性状 (0.11-0.38) 的中等遗传力。估计泌乳时间长度 (0.04-0.10) 和产犊间隔 (0.03-0.06) 的遗传率较低。除了在高产系统下 (0.05) 外，第一次产犊时的年龄估计为中等遗传力 (0.25-0.26)。在生产系统中，泌乳产奶量、305天产奶量和泌乳期长度具有较高的正遗传相关性（0.52-0.96），而泌乳产奶量和泌乳期长度与首次产犊年龄呈负遗传相关性。除低产系统外，产奶量和产犊间隔呈正相关。低和中（0.48 ± 0.20 和 0.46 ± 0.21）和低和高生产系统（0.74 ± 0.15 和 0.）之间泌乳产奶量和 305 天产奶量的遗传相关性。62 ± 0.17) 显着低于 1。因此，低生产系统中的牛奶产量是一种遗传上不同的性状。大多数产奶量和繁殖性状的三个生产系统之间的低遗传相关性表明应根据目标生产系统中的后代性能选择公牛。