Relative effectiveness of Best Linear Unbiased Prediction-animal model vis-Ã -vis other sire evaluation methods for genetic evaluation of Murrah sires

Abstract

First lactation milk production records of 965 Murrah buffaloes sired by 98 bulls were analyzed to estimate breeding values of sires using best linear unbiased prediction animal model (BLUP-AM), best linear unbiased prediction sire model (BLUP-SM), simple regressed least squares (SRLS) and least squares methods (LSQ) of sire evaluation. The effectiveness of the methods was judged using various criteria, viz. error variance, coefficient of determination, coefficient of variation and rank correlation. The highest (2428.02 kg) and lowest (1274.62 kg) average breeding values were obtained by LSQ method. BLUP-AM had the lowest error variance and highest (44.97%) R² value and was adjudged as the most efficient and accurate method compared with others. The CV values of the population by different methods of sire evaluation were more or less similar except BLUP-AM method. The values ranged from 23.59% in BLUP-AM to 26.03 in SRLS method. The rank correlations between expected breeding values for first lactation 305-day milk yield by BLUP-SM and BLUP-AM were the highest (0.94 ± 0.001). The rank correlations of LSQ with either of BLUP-SM or BLUP-AM were almost close and more than 0.75. The rank correlations of BLUP-SM with SRLS and SRLS with BLUP-AM were found to be 0.69 and 0.62. The rank correlations of LSQ with SRLS were found to 0.78. It was concluded based on all the criteria to compare the effectiveness of different methods that BLUP-AM method was more efficient and accurate compared to other methods owing to lowest error variance and highest R²-value.