Next-generation Genomic Selection models

The overall aim of this theme is to develop models that efficiently estimate genomic breeding values, that are needed to perform genomic selection.

In all species, 50k SNP chips are now commonly used. In dairy cattle, a high-density SNP chip with 777k SNP chips is available. Large reference populations (>10,000 animals) now have (imputed) genotypes available for a few hundred thousand SNPs. With the costs of sequencing dropping rapidly, the numbers per animal may increase to >10,000,000. One important aim of this project therefore is to develop models that are able to utilize data with these ever-growing dimensions.

Genomic selection has thus far been applied mainly within breeds. Many industry partners are interested to apply across-breed or across-line genomic prediction, since it may increase prediction accuracy for numerically smaller breeds or lines. Another important aim within this theme is therefore to develop models that enable such across-breed and across-line genomic predictions.