A Comparison of Different Methods of Sire Evaluation for Production and Reproduction Traits of Murrah Graded Buffaloes

Abstract

First lactation monthly test day records of 2329 buffaloes sired by 79 sires distributed in 47 villages under field progeny testing programme in Anand milk shed were used to evaluate sires for age at first calving (AFC), first lactation predicted milk yield (FLY305) and milk fat percent (fat%) traits. The objective of the study was to estimate the sire breeding value and to compare various methods of sire evaluation viz. Simple daughters average, Contemporary comparison, Least squares, Best linear unbiased prediction and Average information restricted maximum likelihood in terms of accuracy and efficiency. Heritability estimates by LS-ANOVA method were 0.252 ± 0.044, 0.065 ± 0.032and 0.073 ± 0.035, respectively for AFC, FMY305 and Fat%. The genetic correlation of AFC with FLY305 and Fat% were -0.782 ± 0.192 and -0.451, respectively while the same between FLY305 and fat percent was 1.044 ± 0.39. The phenotypic correlations of FLY305 with Fat % and AFC were -0.042 and 0.109, respectively while same between Fat % and AFC was -0.099. For AFC, the AIREML had the lowest error variance than LS and BLUP, thus was the most efficient method. For FMY305 and Fat%, the BLUP was found to have lowest error variance and therefore, was the most efficient method. The relative accuracies in terms of rank correlations of D, CC, LS and BLUP methods with most efficient AIREML method were 0.35, 0.90, 0.37 and 0.97, respectively for AFC. These estimates of rank correlation were highly significant (P < 0.01). For FLY305, the value of rank correlations of D, CC, LS and AIREML with most efficient BLUP method were 0.71, 0.71, 0.72 and 0.98, respectively. These estimates of rank correlation were highly significant (P < 0.01). The value of rank correlations of D, CC, LS and AIREML with most efficient BLUP method were 0.89, 0.89, 0.95 and 0.99, respectively and these estimates of rank correlation were highly significant (P < 0.01).