ESTIMATING BODY WEIGHT FROM SEVERAL BODY MEASUREMENTS IN HARNAI SHEEP WITHOUT MULTICOLLINEARITY PROBLEM
M. A. Khan, M. M. Tariq, E. Eyduran*, A. Tatliyer**, M. Rafeeq, F. Abbas, N. Rashid, M. A. Awan and K. Javed***
*Center for Advanced studies in Vaccinology and Biotechnology (CASVAB), University of Balochistan, Quetta, Pakistan
*Biometry Genetics Unit, Department of Animal Science, Faculty of Agriculture, Iğdır University, 76000, Iğdır-Turkiye
**Biometry Genetics Unit, Department of Animal Science, Faculty of Agriculture, Süleyman Demirel University, 32000, Isparta-Turkiye
***Department of Livestock Production,University of Veterinary and Animal Sciences, Lahore, Pakistan
Corresponding Auther: tariqkianiraja@hotmail.com
ABSTRACT
The aim of this study was to estimate body weight from several linear body characteristics collected from 730 Harnai sheep, in Pakistan. For this aim, morphological characteristics viz. body length (BL), withers height (WH), chest girth (CG), paunch girth (PG), face length (FL), length between ears (LBE), length of ears (EARL), width (FTW) and length (FTL) of tail to be useful for breeding purposes were measured. With removing multicollinearity problem, the complex relationship between BW and the measured characteristics was assessed by using scores derived from factor and principal component analyses in multiple regression analysis (MLRA) for male and female sheep. Body weight from morphological characteristics was predicted by Regression tree method. R2 (%), adjusted R2(%), and RMSE values for weight prediction were estimated very high for MLRA (90.6, 90.3, and 4.635 for male sheep, and 92.4, 92.3, and 4.102 for female sheep), whereas use of factor scores in MLRA (87.8, 87.6, 0.352 for male sheep and 92.0, 91.9, and 0.284 for female sheep), and principal component scores (85.9, 85.8, and 0.367 for male sheep and 88.8, 88.7, and 0.335 for female sheep) in MLRA exactly removed multicollinearity problem. Regression tree method explained 84.4 % of total variation in BW. Consequently, use of factor and principal component scores in MLRA gave breeders a good chance without multicollinearity problem for a more accurate estimation of body weight in Harnai sheep compared to the results of MLRA. The results might present valuable knowledge for genetically improving body weight of Harnai sheep.
Key words: Body weight, Factor Analysis, Harnai Sheep, Multicollinearity, Principal Component Analysis, Regression Tree.
|