GENOMIC SELECTION: WHAT IS IT AND HOW DOES IT WORK

This overview was kindly provided by Prof Esté van Marle-Köster of the University of Pretoria.

Genomics defined in broad terms:

Genomics defined in broad terms refers to all DNA-information that can be derived from the DNA of the specific species. Genomics include structural genomics focusing on the discovery of the DNA markers and finding their locality on the genome map, while functional genomics focuses on the expression of the genes that determines the specific functions within the body (for example growth, fertility). Genomic information has a wide application from  identification of genes for genetic defects, parentage testing, finding quantitative trait loci (the traits with economic impact) and providing us with the direct genomic value (info from SNP markers) of an individual that can be applied in Genomic selection.

Brief History:

Genomics is a new field of research and has evolved with the development of molecular technology over the past two decades. DNA technology started out with mapping of DNA markers on chromosomes that was useful for finding genetic defects and DNA based parentage analyses. The sequencing of whole genomes led to the discovery of SNP markers that are abundant markers and these are now the markers that provide us with the information/data for genomic selection.

Genomic selection (GS):

Genomic selection is a concept based on using DNA information to obtain a direct genomic value of the animal. This value based on the SNP marker information must then be “associated“ with the quantitative traits (milk production, fertility, etc.) in order to incorporate the genomic value into the breeding value. Breeders then can receive a genomic enhanced breeding value based on the performance data recorded on farm and the DNA info (genomic info) of the individual animal. GS has therefore the potential of increasing the accuracy of the value (GEBV, genetic index) used to select the specific animal. It provides the opportunity to select animals at a younger age making it unnecessary to wait for progeny tests that take time and are expensive.

Genomic selection, however, cannot and will not replace animal recording, as measurements are essential for interpretation of genomic information. GS has the most potential for traits with low heritability (fertility, longevity traits), sex-limited traits (milk production) or traits difficult to measure (mastitis).

Prerequisites for implementation of GS:

Cattle populations differ with regard to phenotypic traits that can be observed and measured. Over many centuries we managed to select breeds and select for certain traits within a breed, for improvement of the desired traits and characteristics. In selection based on phenotypic variation, all variation is not captured. Research has shown that we still observe more genetic variation within a breed than between breeds, provided that we do not deliberately inbreed for a period of time. Genetic variation does not only differ between breeds, but also between populations, implying that a population of SA Jerseys will differ from populations of Jerseys elsewhere in the world. The dairy breeds are to a certain extent in a good position due to AI practised for many years and therefore genetic linkage among the dairy populations in the world has been established.

In GS the aim is to capture the specific gene (allelic variations) across the whole genome that can be associated with the phenotypic traits in that population. The allelic variations are established by using SNP markers that now are commercially available in high and low density SNP bead chips. These SNP chips were developed for specific populations (training or reference population) and validated within association studies to “link” the SNP variation with the trait of interest in these populations.

Due to the selection in the specific population, research to date has clearly shown that for a breed/population to use GS, a reference population is firstly needed to capture the individual variation in the population and to associate the SNP (sometimes referred to as SNP key) with traits in that population. The selection of the reference animals should be according to available guidelines, in order to cover the variation of the population in full. The first step is, however, to commit to setting-up the reference population.  This, unfortunately, makes the establishment of GS for any breed quite an expensive process. Dairy populations have, compared to beef cattle, the advantage, that high genetic linkage exists due to the extensive use of global AI sires by all dairy countries of the world. This makes genotype profile sharing possible across countries. In Europe and North America dairy breed organizations and service providers have established formal collaborations to share genotypes and ensure a more cost effective process.