Use of simultaneous curve fitting and a four-parameter logistic model to evaluate the nutritional quality of protein sources at growth rates of rats from maintenance to maximum gain.

A four-parameter logistic model was used to describe the dose-response relationship of rats fed diets containing 12 levels of casein, peanut meal or wheat gluten. The model was capable of accurately describing the entire response curve of rats fed diets containing each of the three protein sources. Incorporation of a technique known as parameter sharing into the curve-fitting process facilitated convergence of the parameter estimates for b (the response of rats fed a protein-free diet) and Rmax (maximum response) for all curves when compared with the values observed experimentally. Parameter sharing also provided a method by which the curves could be differentiated on a statistical basis. These data indicate that the relative value of a protein source is dependent on the concentration of the protein in the diet. The application of nonlinear models combined with parameter sharing provides a technique by which protein values can be evaluated at levels of animal response from maintenance to maximum growth.