A data-driven approach to assess general resilience in lambs based on variability in cumulative feed intake

Animal resilience is a key concept in livestock production, as more resilient animals are better capable of coping with short-term environmental challenges. As resilience is not a directly measured trait, many efforts have been made over the past few years to develop resilience indicators that capture the animals' response to perturbations and may become selection criteria in breeding programmes. In this context, the objectives of this study were to (1) derive lamb resilience indicators based on variability in cumulative feed intake and identify a reduced subset of the most promising indicators, (2) evaluate potential differences in resilience phenotypes between sheep breeds and sexes, and (3) investigate the phenotypic correlation between resilience indicators and performance traits in Texel, Corriedale, and Creole sheep breeds. High-frequency individual feed intake data from 436 lambs (176.2 ± 47.4 days of age) from three breeds (Texel, Corriedale, and Creole) were collected during seven feed efficiency trials, which lasted for 53.1 ± 6.0 days on average. These data were used to estimate target trajectories (in the absence of perturbations) of cumulative feed intake using a smoothing function. Deviations between the observed and predicted cumulative feed intake trajectories were associated with potential perturbations and subsequently used to derive eleven resilience parameters. The resilience indicators derived capture different aspects of the perturbation shape, including its amplitude, variability, temporal dynamics, and recovery patterns. Animals had an average of 1.89 ± 0.91 perturbations ranging from one to five across individuals. Differences in estimated marginal means were found for sex within the Texel breed, where males were generally more resilient. Additionally, differences were found between the three breeds within the male group; however, no clear overall superiority was identified. Unfavourable phenotypic correlations were observed between resilience indicators and production traits such as average daily gain and rib eye area, whereas a favourable correlation was observed between resilience and fat thickness. These results indicate that this data-driven approach is a useful tool to evaluate animals' trajectory and responses to unknown environmental challenges. The phenotypic correlation results highlight an anticipated pattern of potential trade-offs between resilience and production that warrants further investigation, including estimation of genetic parameters and biological validation of the derived resilience indicators to enable genetic selection for improved performance and resilience in sheep populations.