A dynamic model as a tool to describe the variability of lifetime body weight trajectories in livestock females
Résumé
Until now, the development of precision livestock farming has been largely based on data acquisition automation. The future challenge is to develop interpretative tools to capitalize on high-throughput raw data and to generate benchmarks for phenotypic traits. We developed a dynamic model of female BW that converts BW time series into a vector of biologically meaningful parameters. The model is based on a first submodel that split a female's weight into elementary mass changes related to biological functions: growth (G component), reserves balance (R component), uterine load (U component), and maternal investment (M component). These elementary weight components are linked to the second submodel, which represents the litter developmental stages (oocyte, fetus, neonate, and juvenile) that drive elementary components of dam weight over each reproductive cycle. The so-called GRUM model is based on ordinary differential equations and laws of mass action. Input data are BW measures, age, and litter weight at birth for each parturition. Outputs of the fitting procedure are a vector of parameters related to each GRUM component and indexed by reproductive cycle. We illustrated the potential application of the model with a case study including growth and successive lactations (n = 202) from 45 dairy goats from the Alpine (n = 27) and Saanen (n = 18) breeds. The fitting procedure converged for all individuals, including goats that went through extended lactations. We analyzed the fitted parameters to quantify breed and parity effects over 4 reproductive cycles. We found significant differences between breeds regarding gestation components (fetal growth and reserves balance). We also found significant differences among reproductive cycles for reserves balance. Although these findings are based on a small sample, they illustrate how use the model can be to adapt herd management and implement grouping strategies to account for individual variability.