Estimation of additive and non-additive genetic variance component for growth traits in Adani goats

• Published in: Tropical animal health and production 2020-03, Vol.52 (2), p.733-742
• Main Authors: Sadeghi, Seyed Abu Taleb, Rokouei, Mohammad, Valleh, Mehdi Vafaye, Abbasi, Mokhtar Ali, Faraji-Arough, Hadi
• Format: Article
• Language: English
• Subjects: Accuracy; Animal models; Bayesian analysis; Biomedical and Life Sciences; Birth weight; Breeding; Dominance; Environmental effects; Epistasis; Error reduction; Genetic diversity; Genetic effects; Genetic variance; Goats; Heritability; Life Sciences; Mathematical models; Model accuracy; Parameter estimation; Phenotypic variations; Regular Articles; Veterinary Medicine/Veterinary Science; Weaning; Weight; Zoology
• ISSN: 0049-4747; 1573-7438
• DOI: 10.1007/s11250-019-02064-0

Non-additive genetic effects are important to increase the accuracy of estimating genetic parameters for growth traits. The aim of this study was to estimate genetic parameters and variance components, specially dominance and epistasis genetic effects, for growth traits (birth weight (BW), weaning weight (WW), 3 (W3), 6 (W6), 9 (W9), and 12 (W12) month weight) in Adani goats. Analyses were carried out using Bayesian method via Gibbs sampler animal model by fitting of 18 different models. All fixed effects (sex, type of birth, age of dam, and year) showed significant effects on BW, WW, W3, and W6, whereas the type of birth and age of dam were not significant on W9 and W12. With the best model, direct heritability estimates were 0.347, 0.178, 0.158, 0.359, 0.278, and 0.281 for BW, WW, W3, W6, W9, and W12 traits, respectively. Maternal permanent environmental effect was significant for BW and WW, but maternal genetic effect was significant only for W3. Dominance and epitasis effects were significant almost for all traits and as a proportion of phenotypic variance were ranged from 0.115 to 0.258 and 0.107 to 0.218, respectively. The range of accuracy of breeding values estimated for growth traits with appropriate evaluation models was from 0.521 to 0.652, 0.616 to 0.694, and 0.548 to 0.684 for the all animals, 10% of the best males and 50% of the best females, respectively. When dominance and epistasis effects added to models, the error variance was reduced and the accuracy of estimated breeding values increased. The accuracy of the best model showed a significant difference with the accuracy of other models ( p