Single-step genetic evaluations: General remarks, background information and necessity

Earliest multi-step genomic evaluations mostly based on genotyped elite animals, especially on so-called sire training sets. The exclusion of animals from commercial farms or of a random sampling from the broad population implies biased genetic evaluations due to violations of general BLUP assumptions. Violations address the effects of pre-selection for genotyping as well as "double counting" of pedigree relationships to estimate conventional breeding values and afterwards of genomic relationships of same animals in the second step of genomic predictions. Prof. Misztal and his group, University of Georgia, US, developed a single-step approach, which bases on individual phenotypes, and on simultaneous and balanced consideration of pedigree and genetic relationships through a so-called H-matrix. The approach generates breeding values for genotyped and non-genotyped animals in one single run. The single-step genetic evaluation is the first choice, especially for novel traits representing only a small sample of genotyped animals with highly accurate conventional breeding values. Single-step modelling implies an increased accuracy of estimated breeding values with a broadening of the selection pool, allowing an intensification of intra-herd selection strategies.