Redundancy Elimination from Morpho-Stuctures of Nigerian Uda Rams Using Principal Component Analysis

Sixteen metric traits; wither height, rump height, neck length, body length, foreleg length, hind-leg length, scrotal length, ear length, tail length, face length, neck circumference, abdominal circumference, scrotal circumference and top-line were measured on 100 randomly selected Uda rams of 6 age (milk tooth, 2-tooth, 4- tooth, 6-tooth, 8-tooth and worn teeth) groups, with the aim of eliminating redundancy using Principal Component Analysis. Mean±SD of all the measured traits were significant (P<0.05) across the different age groups. Pearson’s coefficients of correlation among the measured traits were all highly significant (P<0.01) and ranged from 0.9–0.63. Using varimax rotation of the transformation matrix, two Principal Components (PC) were extracted which total 89.14% of the total variance. PC1 accounted for 86.49% of the total observed variance and was loaded by; wither height, rump height, neck length, foreleg length, hind-leg length, scrotal length, ear length, tail length and face length which describes the overall body size. While PC2 contributes 2.64% of the total observed variance and had its loading (rump length, body length, neck circumference, abdominal circumference, scrotal circumference and top-line) for meatiness. The results suggest that PCA could be explored in breeding programs with greater success to eliminate variable redundancy.