主成分和回归分析在安格斯牛选育中的应用

Application of principal component and regression analysis in Angus cattle breeding

【目的】揭示安格斯牛的体型外貌特征和生长发育规律,分析各性状参数间的相关性并探求准确而简易的体质量估测方法。【方法】以宁夏地区1753头安格斯初生牛、1764头安格斯断奶牛和1642头安格斯周岁牛为对象,对其3个生长阶段的6个体尺指标和体质量进行主成分分析和逐步多元回归分析。【结果】安格斯牛不同生长阶段各性状间均存在极显著(P<0.01)相关性。由主成分分析结果可知,安格斯牛初生阶段的体质量与体尺指标可综合为4个主成分指标,分别为高度因子、围度因子、管围因子和体长因子,累计贡献率为89.5%;断奶和周岁阶段均分别筛选出3个主成分,其中第1主成分和第2主成分均为体型外貌因子和高度因子,第3主成分分别为体长因子和体质量因子,累计贡献率分别为93.1%和88.1%。通过构建安格斯牛不同生长阶段估测体质量的回归模型,得到初生、断奶、周岁3个阶段的最优回归方程,其决定系数R2分别为0.8170,0.9563和0.5511,表明各阶段体质量与体尺性状间存在明显的线性关系。【结论】不同阶段的主成分反映了安格斯牛不同阶段的生长发育特征,各阶段体质量回归方程的决定系数表明各回归方程存在一定的参考价值。

【Objective】This study revealed the physical appearance,growth and development of Angus cattle,analyzed the correlation between trait parameters and established an accurate and simple way to estimate their body mass.【Method】The principal component analysis and stepwise multiple regression analysis were performed on 6 body size indexes and body mass of 1753 newborn Angus,1764 weaned Angus and 1642 aged Angus in Ningxia.【Result】There was extremely significant(P<0.01)correlation between the traits of Angus cattle at different growth stages.According to principal component analysis,body mass and body size can be integrated into four principal component indexes of height factor,circumference factor,tube circumference factor and body length factor with a cumulative contribution rate of 89.5%.Three principal components were screened in the weaning state and first year.The first principal component of body appearance factor and the second principal component of height factor were same and the third principal components were body length factor and body mass factor,respectively.The cumulative contribution rates were 93.1%and 88.1%,respectively.By constructing regression models for estimating body mass at different growth stages of Angus cattle,the optimal regression equations of initial stage,weaning stage and the first year were obtained with determination coefficients(R2)of 0.8170,0.9563 and 0.5511,respectively.The relationship between body mass and body size traits at each stage was linear.【Conclusion】The principal components of different stages reflected the growth and development characteristics of Angus cattle at different stages.The determination coefficients indicate that the regression equations were valuable.