A novel modelling approach to quantify the response of dairy goats to a high-concentrate diet

High-producing ruminants need high-concentrate diets to satisfy their nutrient requirements and meet performance objectives. However, such diets induce sub-acute ruminal acidosis (SARA), which will adversely affect dry matter intake and lead to lower production performance. This work develops a novel modelling approach to quantify the capacity of dairy goats to adapt to a high-concentrate diet challenge at the individual level. The animal model used was dairy goats (from Saanen or Alpine breed), and rumen pH was used as the indicator of the response. A three-step modelling procedure was developed to quantify daily scores and produce a single global index for animals' adaptive response to the new diet. The first step summarizes the post-prandial kinetics of rumen acid status using three synthetic variables. In the second step, the effect of time on the response of goats is described, in the short and long terms. In the last step, a metric based on phase trajectories ranks goats for their resilience capacity. This modelling procedure showed a high variability among the goats in response to the new diet, highlighting in particular their daily and general strategies to buffer the effect of the diet change. Two main categories of adaptive strategies were observed: (i) acid status increased, but the goats tried to minimize its variations, and (ii) acid status oscillated between increases and decreases. Such phenotyping, alongside other behavioral, digestive, and metabolic measures, can help to determine biomarkers of goats' capacity to adapt to diets of higher nutritive value and to increase production performance without compromising their health status. Quantifying the capacity of goats to buffer the effect of highly fermentable diets helps to better adapt feed to animals in precision livestock farming. This procedure is generic and can be adapted to any indicator of animal health and performance. In particular, several indicators can be combined to assess multi-performance, which is of major interest in the context of selection for robust animals.