Adding value to test-day data by using modified best prediction method

Computation of lactation yields from test-day data has lost much of its importance for genetic evaluations as the use of test-day models is currently quite widespread. In the other hand its interest for intra-farm management is increasing as a base for advanced management tools. The first and principal aim of this study was to develop a method which takes into account advantages and disadvantages of existing methods, and to test its potential to provide useful management tools to dairy farmers. A test-day model with modifications to able daily run and management tools was developed. Because of its similarities with best prediction, the method developed here was called modified best prediction. The second objective was to compare the accuracy of this new method with best prediction and test interval methods. Modified best prediction showed good results for predicting daily yields and was slightly better than best prediction for lactation yields prediction. Management tools obtained with modified best prediction are explained.