Implementation of Genomic Selection conceptually proceeds in two steps:
(1) Estimation of the effects of chromosome segments in a reference population and
(2)Prediction of genomic EBVs(GEBVs) for animals not in the reference population
The difficulty in step1 is that a very large number of haplotype effects across the chromosome segments must be estimated, most likely from a dataset where the number of phenotypic observations is less than the number of chromosome segemnt effects to be estimated.