Comparison of models involving direct and maternal genetic effects in sheep

The objective of this study was to estimate the genetic parameters and variance components associated with the 90-day live weight of Central Anatolian Merino lambs, utilising six distinct animal models. The research focused on the identification of direct and maternal genetic effects, as well as the genetic covariance between these effects. The study material comprised data collected in 2024 from the Bahri Dağdaş International Agricultural Research Institute, which is under the jurisdiction of the Ministry of Agriculture and Forestry of the Republic of Türkiye. The data set included information on 1693 lambs born to 49 rams and 706 ewes. The data analysis was conducted utilising the BUGA software. In the modelling process, Model 1 included only the direct genetic effect; Model 2 included both direct genetic and maternal environmental effects; Model 3 included direct and maternal genetic effects; Model 4 added the direct x maternal genetic covariance; Model 5 additionally included the maternal permanent environmental effect; and Model 6 was structured to include all of these variance components simultaneously. The results indicate that direct genetic variance ranged from 1.793 to 3.5674, maternal genetic variance from 1.415 to 4.005, residual variance from 11.257 to 13.542, and phenotypic variance from 17.056 to 17.109. The estimates of direct heritability ranged from 0.105 to 0.208, and maternal heritability ranged from 0.083 to 0.234. The estimation of the covariance between direct and maternal genetic effects was conducted in Models 4 and 6, yielding values of 1.533 and 0.477, respectively. These values correspond to genetic correlations of 0.728 and 0.299, respectively. Furthermore, the maternal permanent environmental variance was found to contribute between 2.112 and 4.005 units to the phenotypic variance, thus highlighting the significant role of environmental factors in determining live weight. Following a thorough examination of the models in question, it was determined that Model 1 was the most suitable model based on the Akaike Information Criterion (AIC). Conversely, Model 6 was deemed to be the optimal model according to the Bayesian Information Criterion (BIC). The log-likelihood (LogL) values indicated that Models 2 through 6 exhibited a superior overall fit. In conclusion, the incorporation of maternal genetic and permanent environmental effects into the models enhanced the accuracy of genetic parameter estimation and facilitated a more realistic decomposition of variance components. These findings contribute to a deeper understanding of the interaction between genetic and environmental factors, and provide a solid scientific foundation for improving the effectiveness of selection programmes in sheep breeding. KEYWORDS: Birth weight, Variance components, Genetic parameter, Genetic effect, Maternal effect