Predicting body weights of Kolhapuri sheep using principal component scores

This study, in addition to defining the morphometric structure of Kolhapuri sheep, predicts body weights using the principal component analysis (PCA) of the morphometric traits. Data on body weights and 13 body measurements were collected on 265 Kolhapuri sheep in the breeding tract. For the traits measured, the coefficients of variation (%) ranged from 5.1 (rump width) to 18.8 (body weight). In general, the phenotypic correlations between body weights and morphometric traits were positive and significant (p<0.05/0.01). Very high correlation (0.96) between withers height and rump height suggested colinearity between the two traits. Two components were derived from the PCA of morphometric traits with the total variance explained in the data by 67.4%. Using morphometric variables as predictors, regression models explained up to 85% of the body weight variance, while up to 86% was explained by using principal components as predictors. The multiple linear regression analysis suggested that use of the principal component scores was better than the use of original correlated traits in predicting the body weights of sheep.