RAJA, K.N.ASHISH YADAV2023-12-052023-12-052022https://krishikosh.egranth.ac.in/handle/1/5810202205The aim of the present study was to genetically evaluate Murrah buffaloes for various production traits recorded throughout their lifetimes. The study relied on lactation records of Murrah buffalo, maintained at the record room (AGB Division) of ICAR-NDRI, Karnal from the period 1971 to 2020. Through Bayesian analysis utilising the BLUPF90 family of programmes, genetic analysis was carried out using animal models with a combination of random factors, such as the direct additive genetic effect of the animal, the direct maternal genetic effect, and the maternal permanent environmental effect. Using the most appropriate models (based on DIC value) that account for direct and maternal influences fitted to the majority of the traits, (co)variance components and genetic parameters were estimated. The heritability estimates of FL305MY, TMY, LL, PY, PL and 305 MYAP in the most appropriate models were 0.49 ± 0.007, 0.20 ± 0.011, 0.04 ± 0.001, 0.10 ± 0.001, 0.10 ± 0.009 and 0.21 ± 0.003 respectively with repeatability estimates of TMY, LL, PY, and 305MYAP as 0.40, 0.20, 0.21 and 0.34. In comparison to the animal model with only additive genetic effects of the animal in each trait, lower error variance and higher additive variance were found for production and lifetime traits through various models based on partition of variance, indicating that it is crucial to use the best statistical model for elucidating genetic parameters that incorporate suitable random and fixed effects. The comparison of sires based on their breeding value (obtained from AIREML as well as the Bayesian approach) and ranking them, the correlation of which obtained using spearman’s rank correlation revealed that both approaches provide similar results as these all were normally distributed traits, with the best BV estimates for FL305MY, TMY, LL, PY, PL and 305 MYAP as 404.40, 415.99, 20.43, 1.39, 221.89 and 398.19 respectively using Bayesian approach. AIREML revealed estimates of BV of the given traits as 401.22, 409.02, 18.11, 1.37, 111.82, and 393.55. In terms of the accuracy of breeding values based on the standard error predicted, the Bayesian method provides a slightly better outcome with relative efficiency over the AIREML method as 0.73 %, 1.2 %, 7.2 %, 1.2 %, 42.94 %, and 0.78 % in respective traits. It is recommended to use the Bayesian method to estimate the breeding values of Murrah sires for production traits, which in turn can be used for ranking and selection of bulls.EnglishThesis