Genetic Heterogeneity of Residual Variance for Foot Score Traits in American Angus Cattle

Foot conformation is one of the main breeding goals in recent beef cattle breeding programs because it directly affects productivity, economic losses, animal welfare and longevity. Genetic heterogeneity of residual variance can be used to improve production uniformity in animal breeding programs because recent studies have shown that residual variance is partially under genetic control, allowing reduction of variability through selection. Despite being an important breeding goal, research on genetic heterogeneity of residual variance for conformation traits, such as foot angle and foot claw, is still scarce in livestock species. The objectives of our study were (1) to investigate the extent of genetic heterogeneity of residual variances on two conformation traits: foot angle (FA) and claw set (CS) in Angus cattle using genetic homogeneity (M1) and two genetic heterogeneity of residual variance models, including a double hierarchical generalised linear model (DHGLM, M2) and a genetically structured environmental variance model (M3). Genetic parameters for means and residual variances were estimated using M2 and M3. The dataset included 45,667 phenotypic records for FA and CS (scores from 1 to 9 with 5 being ideal) of American Angus cattle recorded from 2009 to 2021. M1 and M2 were fitted using average information restricted maximum likelihood, and M3 was fitted using Markov chain Monte Carlo. Heritability estimates for the means of FA (0.19 ± 0.007 for M1, 0.11 ± 0.005 for M2 and 0.09 ± 0.003) and CS (0.16 ± 0.005 for M1, 0.10 ± 0.004 for M2 and 0.08 ± 0.03) were within the range reported in the literature, but M2 and M3 estimates were lower than M1. Genetic heterogeneity of residual variance was assessed using three parameters: heritability for residual variance, genetic coefficient of variation, and correlation between mean and residual variance. Although heritability estimates for residual variance in M2 were low (0.08 for FA and 0.001 for CS), our results suggest that residual variance is partially under genetic control. The genetic coefficients of variation estimates were 0.08 (M2) and 0.06 (M3) for FA, and 0.06 (M2) and 0.02 (M3) for CS, indicating that selection on the trait mean would also change the residual variance. Our results for FA and CS showed moderate positive genetic correlations in M2 (0.52 for FA and 0.41 for CS) and M3 (0.35 for FA and 0.33 for CS) between mean and residual variance. Positive correlations may limit the response to selection unless other breeding strategies, such as selection indices, are used. FA and CS are promising traits for uniformity or resilience indicators because they are phenotypes that can be collected throughout the production cycle using traditional or digital data recording systems. Our results demonstrate the potential to modulate variability through breeding strategies and present an opportunity to evaluate the uniformity of foot score traits in beef cattle.