Disease & heat resistance

Selection index including resilience measured in controlled environments improves performances in disturbed environments

---

Abstract

Resilience, the capacity to withstand or rapidly recover from disturbances, is essential for sustaining livestock performance under challenges. Some authors suggest that selection programs focused on production traits indirectly capture resilience, assuming animals that maintain high performance under stress are resilient. However, this approach may be less effective when selection candidates are raised in low-disturbance environments (LDE), while commercial animals are in high-disturbance environments (HDE), as in the pig industry. This study aimed to evaluate selection response in HDE, comparing selection on the EBV of the observed production trait (integrating resilience and production), versus selection based on indices combining true production and resilience potentials. We simulated growth trajectories for animals from a simplified pig population spanning 15 non-overlapping generations (18,162 animals). Each trajectory incorporated an elastic response to disturbances. The model integrated a theoretical growth trajectory, simulated using a degree-2 genetic random regression model calibrated for moderate heritability (0.20 to 0.35), and disturbances' effect, moderated by animals' resilience. Resilience was characterized by two unobserved phenotypes on a [0,1] scale, modelled using logit genetic models with an underlying heritability of 0.5. Disturbances, including their starting date, duration, and intensity, were randomly sampled in each environment, with higher probability and intensity range in HDE. The simulation parameters were calibrated using French nucleus farms data. Growth performances in LDE were used to select animals, while body weight at 100 days of age (BW160) in HDE was the phenotype to improve. Animals were selected under two strategies: selection based on EBV of observed BW160, or selection based on indices combining EBVs of theoretical BW160 and resilience. In this second approach, to isolate growth potential from resilience, we excluded longitudinal growth records of animals following a disturbance, detected using UpDown R-package, and fitted a random regression animal model on the longitudinal (truncated) phenotype to predict theoretical BW160. For non-disturbed animals, theoretical BW160 is the observed BW160. Resilience phenotype was derived from the study of the evolutionary dynamics of the growth phenotype of animals facing a detected disturbance. Subsequently, single-trait and multiple-trait animal models were fitted on observed BW160 for the first strategy, and on theoretical BW160 and resilience phenotypes in the second strategy, respectively. Then animals were selected based on the EBV of observed BW160 or combinations of EBVs of theoretical BW160 and resilience. Strategies were evaluated under different genetic correlations (corG) between BW160 and resilience (0, -0.25), based on literature, with 100 replicates each. Both strategies increased observed BW160 across generations in HDE. However, the second strategy achieved the highest selection response when corG was 0 or −0.25, resulting in 5 to 38% higher BW160 after 10 generations. These results demonstrate that dissociating resilience from production can enhance animals' productivity in uncontrolled environments.

Keywords: 2026

How to Cite:

Antonios, S. & David, I., (2026) “Selection index including resilience measured in controlled environments improves performances in disturbed environments”, World Congress on Genetics Applied to Livestock Production Digital Archive 2026(1): 2285364. doi: https://doi.org/10.31274/wcgalp.23657

Rights: 1