Testing BLUP methodology for large white pigs selection

The article examines the convergence of BLUP-based (Best Linear Unbiased Prediction) estimates of economically important traits in Large White pigs using different biometric models. The results indicate that significant differences in model quality, reflected particularly in coefficients of determination, lead to considerable discrepancies in evaluation outcomes. For instance, the highest convergence was observed for the FCR (feed conversion ratio) trait: at the population level, the repeatability coefficient was 0.981, with average rank differences for the 20 most reliably evaluated individuals amounting to ∆ = -19.05. Meanwhile, the difference in the coefficients of determination of the models used was 6.2%. The lowest convergence was observed for the BWG (body weight gain) trait, where the repeatability coefficient was 0.720, and the rank difference reached ∆ = 119.55. The coefficients of determination for the models used to assess this trait differed by 27%, or 1.81 times. Such discrepancies under simulated conditions of limited availability of primary zootechnical accounting data underscore both the necessity of considering the full range of available variables affecting the variability of the evaluated trait and the importance of using the same set of variables to ensure comparability of results.