Best Linear unbiased predictor analysis

This chapter describes the theory and application of best linear unbiased prediction (BLUP) in a hybrid breeding program. The BLUP procedure requires two kinds of information that should be readily available to breeders: performance data on single crosses that have been tested and genetic relationships among the parental inbreds. The principle underlying BLUP is the use of information from relatives. The classic example of using information from relatives is the prediction of milk yield performance of a dairy bull. The BLUP procedure has been used for several years in the North American maize breeding program of Limagrain Genetics, a subsidiary of Groupe Limagrain. The effectiveness of BLUP was evaluated with a cross-validation procedure. The use of both trait and marker data in BLUP is currently being studied. But even without marker information, the BLUP procedure has vast potential for increasing the efficiency of present-day hybrid breeding programs.