西藏河谷型绵羊表型指标主成分分析

Principal Component Analysis of Phenotypic Indices in Tibetan Valley Sheep

为分析西藏河谷型绵羊表型指标主成分,本实验随机选取154只成年西藏河谷型绵羊母羊进行体质量称量、体尺指标测量,利用统计软件对相关变量进行分析、拟合以及预测,评估西藏河谷型藏绵羊的体型。结果表明:体长(57.73±5.04)cm、体高(51.81±4.46)cm、腹围(82.60±7.7)cm、胸围(74.82±7.18)cm、臀高(53.02±3.20)cm、臀长(17.82±1.50)cm、胸深(23.21±2.11)cm、胸宽(5.78±3.18)cm、臀宽(12.48±2.37)cm、管围(6.12±0.45)cm、头宽(10.94±1.41)cm与体质量(26.76±4.62)kg呈极显著正相关(p<0.01)。以13个表型指标进行逐步拟合回归时,统计学上具有意义的指标为腹围、胸宽、胸深、臀高、体长,其中胸宽对体质量的间接作用最大,胸宽、胸深、臀高、体长对体质量的间接作用大于直接作用;基于14个表型指标进行主成分分析,提取的4个成分占总方差65.19%,其中PC1与体型相关指标有较高关联,而PC2与头型指标相关,利用主成分分析提取因子拟合体尺与体质量线性回归方程为BW=26.760+3.957PC1-0.678PC3-0.384PC2(R2=0.762,SE=0.492)。综上,主成分分析可用于评价河谷型藏绵羊成年母羊体型外貌,使用主成分因子预测其体质量比用表型指标直接进行线性回归拟合更准确。

To analyze the principal component of phenotypic indices of Tibetan valley sheep,154 adult ewes were randomly selected,and weight and body size were measured.Relevant variables were analyzed,modeled and predicted with statistic software to access the somatotype of Tibetan valley sheep.The results showed that there was a positive correlation between body length(57.73±5.04)cm,whither height(51.81±4.46)cm,abdominal perimeter(82.60±7.7)cm,heart girth(74.82±7.18)cm,rump height(53.02±3.20)cm,rump length(17.82±1.50)cm,thorax depth(23.21±2.11)cm,chest breadth(5.78±3.18)cm,rump width(12.48±2.37)cm,shin circumference(6.12±0.45)cm,head width(10.94±1.41)cm,and the body weight(26.76±4.62)kg(p<0.01).In stepwise fitting regression with 13 phenotypic indexes,abdominal circumference,chest breadth,thorax depth,rump height,and body length had the statistically significant indexes,among which chest breadth had the most indirect effect on body weight,and the indirect effect on chest breadth,thorax depth,rump height,body length,and body weight was greater than the direct effect.Based on 14 phenotypic indicators,principal component analysis showed that four components were extracted,accounting for 65.18%of the total variance.Among them,PC1 was highly correlated with somatotype-related indicators,and PC2 was correlated with headform-related indicators.Using principal component analysis to extract factors,the linear regression equation of body size and weight was fitted,BW=26.760+3.957PC1-0.678PC3-0.384PC2(R~2=0.762,SE=0.492).Taken together,principal component analysis can be used to evaluate the body appearance of adult ewes of Tibetan valley sheep,and it is more accurate to predict body weight by using principal component factors than by using phenotypic indexes directly to perform linear regression fitting.