Do more mechanistic models increase accuracy of prediction of metabolisable protein supply in ruminants?

Ruminal microbes partially degrade dietary protein and synthesise microbial protein, which, along with undegraded true protein, contributes to metabolisable protein for the animal. Rumen models have been developed over the past several decades in an effort to better predict metabolisable protein supply for ration formulation for ruminants. These models have both empirical and mechanistic components. Separation of dietary protein into fractions that include non-protein nitrogen, true protein and unavailable protein has been a fundamental element of these models. Ruminal degradation of one or more true protein fractions is then estimated on the basis of the kinetics of digestion and passage. Some models use the same method to predict substrate supply for microbial protein production. Although mechanistic models have been extensively used in diet-formulation programs worldwide, their ability to improve accuracy of prediction of metabolisable protein over simpler empirical models is questionable. This article will address the potential of mechanistic models to better predict metabolisable protein supply in ruminants as well as their limitations. Metabolisable protein in ruminants is derived from dietary protein that is not degraded in the rumen and protein synthesised by rumen microbes that passes to the duodenum. Mechanistic rumen models have been developed in an attempt to improve the prediction of metabolisable protein for ration formulation. These models are more complex than are simple empirical models and their accuracy is limited by the availability of data to parameterise them. This article discusses the limitations of mechanistic models for the prediction of metabolisable protein in ruminants. [ABSTRACT FROM AUTHOR]