TY - JOUR ID - 2652 TI - Genetic Properties of Some Economic Traits in Isfahan Native Fowl Using Bayesian and REML Methods JO - Poultry Science Journal JA - PSJ LA - en SN - 2345-6604 AU - Salehinasab, M AU - Latifi, M AU - Zerehdaran, S AU - Alijani, S AD - Department of Animal Genetic and Breeding, Faculty of Animal Science, Gorgan University of Agricultural Sciences and Natural Resources, Gorgan, Iran. AD - Department of Animal Science, College of Agriculture, Tabriz University, Tabriz, Iran. AD - Department of Animal Science, College of Agriculture, Ferdowsi University of Mashhad, Mashhad, Iran. Y1 - 2015 PY - 2015 VL - 3 IS - 2 SP - 143 EP - 149 KW - Native fowl KW - Maternal effect KW - REML approach KW - Direct heritability KW - Bayesian approach DO - 10.22069/psj.2015.2652 N2 - The objective of the present study was to estimate heritability values for some performance and egg quality traits of native fowl in Isfahan breeding center using REML and Bayesian approaches. The records were about 51521 and 975 for performance and egg quality traits, respectively. At the first step, variance components were estimated for body weight at hatch (BW0), body weight at 8 weeks of age (BW8), weight at sexual maturity (WSM), egg yolk weight (YW), egg Haugh unit and eggshell thickness, via REML approach using ASREML software. At the second step, the same traits were analyzed via Bayesian approach using Gibbs3f90 software. In both approaches six different animal models were applied and the best model was determined using likelihood ratio test (LRT) and deviance information criterion (DIC) for REML and Bayesian approaches, respectively. Heritability estimates for BW0, WSM and ST were the same in both approaches. For BW0, LRT and DIC indexes confirmed that the model consisting maternal genetic, permanent environmental and direct genetic effects was significantly better than other models. For WSM, a model consisting of maternal permanent environmental effect in addition to direct genetic effect was the best. For shell thickness, the basic model consisting direct genetic effect was the best. The results for BW8, YW and Haugh unit, were different between the two approaches. The reason behind this tiny differences was that the convergence could not be achieved for some models in REML approach and thus for these traits the Bayesian approach estimated the variance components more accurately. The results indicated that ignoring maternal effects, overestimates the direct genetic variance and heritability for most of the traits. Also, the Bayesian-based software could take more variance components into account. UR - https://psj.gau.ac.ir/article_2652.html L1 - https://psj.gau.ac.ir/article_2652_a4cafc6407476d845cee13c9666f72f0.pdf ER -