Mapping genomic regions affecting resilience traits in a large dairy farm of Holstein cows

Introduction: This study evaluated the genetic architecture of resilience indicators in Holstein cows managed in a herd equipped with automatic milking systems (AMS) from 2017 to 2024. Methods: Four resilience indicators were calculated based on deviations in daily milk yield: log-transformed variance (LnVar), autocorrelation of residuals (rauto), weighted frequency of perturbations (wfPert), and accumulated milk losses due to perturbations (dPert). Polynomial quantile regression models were applied to 594,481 daily records from 966 cows, with data filtered for completeness and lactation duration. Genome-wide association studies (GWAS) were performed using selective genotyping coupled with DNA pooling statistics. Results: Descriptive statistics revealed that LnVar increased with parity, indicating greater production variability in older cows, while rauto remained stable, suggesting a consistent ability of cows to recover from production perturbations. Both dPert and wfPert increased across lactations, reflecting greater cumulative losses and perturbation frequencies. Genes related to immune response, energy metabolism, and tissue integrity were identified. Discussion: These findings suggest a multifactorial complex genetic nature of resilience and disclose the involvement of several genes that can explain both the physiology related to production and response to stressors.