MODELING NUTRITIONAL SYSTEMS TO OPTIMIZE CATTLE FARMING IN THE STATE OF GOIÁS

Coordinator: Edgar Alain Collao Saenz

Animal production is the result of a fragile balance between anabolic and catabolic processes involved in the transformation of metabolic pools. Mathematical simulation models have been developed in an attempt to describe and better understand these complex physiological processes. Some of these sets of equations represent the dynamics of nutrients in meat and milk production and constitute a valuable instrument in the development and evaluation of research strategies in nutrition. As qualitative knowledge of ruminant metabolism has increased, it has become possible to develop quantitative approaches that allow for the extension of understanding and integration of various aspects of animal nutrition research. There are several types of models used in agricultural production systems. The first models consisted of regression equations between nutrient intake and animal performance (weight gain, milk production, etc.). Subsequently, empirical models were proposed with the intention of predicting the nutritional requirements for a given animal performance based on average daily gain and live weight. Despite their practical application in farm conditions, these models are limited by the small set of experimental data used in their construction. Furthermore, empirical models consider only one level of aggregation, therefore the needs for understanding and predicting animal responses (quantity and quality of products, efficiency, comfort, etc.) in relation to changes in diet and other variables cannot be met by these approximations (Schmidely, 1996). More recently, dynamic mathematical models based on biochemical reactions have been proposed, which not only summarize existing data, but also show gaps in current knowledge and where greater efforts and research should be directed. Several types of models for use in ruminant research have been described. Tedeschi (2019) provides a comprehensive description of the application of models to support decision-making in ruminant nutrition, characterizing different paradigms and approaches used, and briefly describes the evolution of different lines of thought in nutritional modeling. In the long term, the use of models to predict feed utilization would have four main advantages over traditional feeding systems: a) better use of detailed data on the chemical composition of feed; b) consideration of the interaction between energy and protein; c) prediction of milk constituents in lactating females, or the fat-protein ratio in the carcass of growing animals; and d) prediction of responses instead of just calculating requirements (Gill, 1996). The use of simulation models can also substantially reduce the number of physical tests required for the technical and economic evaluation of cattle diets, submitting only those that present the best results in the simulations to field tests. In an attempt to improve the accuracy of these systems, real data should be obtained in the field to serve as a parameter in these systems and improve the diets provided to high-production animals. The use of more precise nutrition strategies, based on specific nutrients or components in feed, has recently grown to optimize their use and excretion into the environment. Precision nutrition can minimize the oversupply of nutrients, which end up being excreted via feces, urine or milk. In general, when some nutrients are no longer limiting in cattle diets, an increase in production is expected. In this project, we intend to experimentally evaluate the performance of cattle when supplemented with different ingredients that may have specific effects on production and use these results to parameterize the animal response in existing simulation models.