Morphological characterization of indigenous goat population in Ethiopia using canonical discriminant analysis

Abstract

This study was conducted to characterize the indigenous goat population in five zones vis-à-vis South Gondar (S-Gondar), Jimma, Sidama, Gamo, and North Shoa (N-Shoa) representing various agroecological zones of Ethiopia using morphometric traits. Seven morphometric traits were measured from 2540 goats of both sex from various age groups ranging from yearling to four years and above. The traits scored were live weight (LW), body length (BL), wither height (WH), heart girth (HG), rump length (RL), rump width (RW) and ear length (EL). A multivariate canonical discriminant analysis in combination with cluster and discriminant analysis was applied to identify the combination of variables that best differentiate the five goat populations based on morphometric traits. Results indicated that the N-Shoa goats demonstrated the largest LW, HG and RL and could be considered as large-framed goat types. The S-Gondar goats had the highest WH describing them as the tallest goat types. Jimma goats had the smallest morphometric measurements and could be classified as small-sized goat types. Gamo goats had the lowest WH with a relatively larger RW being the shortest amongst the studied goat breeds. The cluster analyses showed that goats of Jimma tended to group separately whilst those of N-Shoa and S-Gondar as well as Sidama and Gamo clustered closely. The stepwise discriminant analysis indicated that RL followed by RW were the most discriminating variables to separate the five goat populations. The canonical discriminant analysis determined four canonical variables (CAN) of which CAN1 (RL and HG) and CAN2 (RW, EL, and LW) accounted for 80.2 and 14.3 % of the total variation, respectively. The canonical discriminant analysis further showed that the Gamo goat is distinct from the other four populations. All Mahalanobis distances among goat population were significant (p < 0.001) being the shortest between Sidama and N-Shoa goats (3.92) followed by that of S-Gondar and Jimma (4.17). The largest distance was noted between N-Shoa and Gamo (68.2) followed by that of Sidama and Gamo goats (53.3). Distances among the other zones were intermediate ranging from 10.7 to 16.2. The respective 99.7, 98.7, 98.6, 97.4 and 95.5 %, of the Gamo, Jimma, S-Gondar, N-Shoa and Sidama goats were correctly assigned to their source population with an overall success rate of 98 %. The present work revealed that characterization of indigenous livestock based on morphological traits through the application of multivariate analysis is a viable option in regions where molecular tools are inaccessible provided the sample sizes of animal population are large enough to effectively represent the study regions for reliable results.