Current Status Of Genomic Prediction and GWAS In Livestock

Genomic prediction offers opportunities to improve rates of genetic improvement, but the value proposition varies widely between livestock industries.  There are real benefits from accurate genomic prediction, but large training populations are prerequisite, and validation must be demonstrated before adoption by breeding program managers. The clearest benefits for investing in training populations have been in dairy cattle where conventional programs had required expensive progeny testing of candidate sires.  Nevertheless, industry and public research investments have now resulted in sufficiently large training populations that genomic prediction has also been implemented in beef cattle, pigs, layer and broiler chickens, and sheep.  In contrast to genomic prediction in humans, effective population sizes in livestock are typically much smaller, and the individuals that comprise the targets for prediction can sometimes be closely related to individuals used in training.  Genome-wide association studies (GWAS) of complex traits are of interest to improve scientific understanding of the number and nature of genes influencing performance, and to improve the accuracy of genomic prediction.  Such studies are being undertaken in targeted research populations as well using industry nucleus populations genotyped for genomic prediction.  Accuracy of genomic prediction tends to be much better in simulation studies than has been achieved in industry populations, except in some cases when validation is carried out using immediate offspring of individuals comprising the training populations.  The variety of traits being subjected to genomic analyses, and the breadth of population structures across these industries, is providing unique insight into the circumstances required for accurate prediction.