Predicting the effects of animal variation on growth and food intake in growing pigs using simulation modelling

Abstract

AbstractAll pig nutrition models to date predict growth responses of either an individual animal or the average animal of a given population over time. Translating the predicted nutrient requirements from the average animal to the population introduces a number of errors as the cause-and-effect response of the average animal is different from the population response. To overcome the problem of estimating the requirements for a given population using models it is necessary to simulate a number of individuals representative of a population and then average these results. This approach however, requires a knowledge of those animal characteristics that vary between individuals and the nature of their distribution. In this paper a scaled growth rate constant (B*), protein weight at maturity (Pm) and the ratio of lipid to protein at maturity (LPRm) are the parameters used to define an individual animal. As no data existed from which the nature of the distribution of B*, Pm and LPRm can be estimated for pigs of different strains and sexes, and due to the impracticality of determining this variability by experimentation, a simulation model was used to estimate the variations within each parameter. In addition this paper quantifies the subsequent effects these distributions have on the genetic variability of average daily gains (ADG) and daily food intake (TI) over a live-weight range of 25 to 90 kg. Comparisons were made between the genetic variation determined by modelling and those published in the literature. The results indicated coefficients of variation for B*, Pm and LPRm of 0·01, 0·05 and 0·10, respectively. An increase in the variability of all three parameters resulted in an increase in the variation in ADG whilst only an increase in the variation of B* and LPRm affected the distribution of FI.