Comparative study of sire evaluation methods for breeding value estimation in Jersey x Sahiwal crossbred sires under field conditions

This study utilized 4,877 first lactation records of Jersey x Sahiwal crossbred cattle, under field progeny testing programme across 24 mandals in Chittoor district, Andhra Pradesh.Breeding values of sires were estimated using first lactation 305-day milk yield data through three methods: Least Squares Means (LSM), Best Linear Unbiased Prediction (BLUP) and Derivative-Free Restricted Maximum Likelihood (DFREML).Given the variability in the efficiency of these methods, the study aimed to assess their effectiveness in accurately predicting the breeding values of Jersey x Sahiwal sires.The average breeding values for 176 sires were 2249.26kg (LSM), 2242.87 kg (BLUP) and 2242.87 kg (DFREML).While LSM showed the highest average value, DFREML displayed the largest range between maximum and minimum breeding values (1368.57kg), indicating its superior ability to differentiate among sires.The methods were compared using critical evaluation parameters, including within-sire variance (error variance), coefficient of determination (%), coefficient of variation (%) and rank correlations.DFREML emerged as the most precise method with the lowest error variance (160,916.26kg) and the highest coefficient of determination (66.53%), reflecting its accuracy.Moreover, DFREML showed the smallest CV (26.71%), signifying its stability, while BLUP was the least stable with a CV of 32.80%.High rank correlation between LSM and DFREML indicated that either method could be used to achieve similar sire rankings.However, considering its superior precision, accuracy and stability, DFREML proved to be the most efficient and accurate method, making it the preferred choice for sire evaluation in Jersey x Sahiwal crossbred cattle.