The Basics of Genomics

The traditional method of breeding value estimation depends on the collection of field data like milk recording, type classification, fertility data etc. This results in reliable genetic evaluations but it is a time consuming process. Sires of cows are 4 to 5 years old at the moment that their breeding value has a reliability of 80% or more and dairymen can decide to use these bulls in their herds. Until that moment these bulls are not available and only 10 to 15% become available as proven product in the market. You could say that we are looking at a black box, we can see the outside (the phenotypic results from the field data) and are estimating how this is influenced by the DNA of the animals.

Genomics now opens the black box to us, because now we can really link the DNA of the animals to their phenotypic results. DNA of bulls and cows is dispersed over chromosomes in each cell of its body. The building blocks of DNA are four kinds of bases connected in pairs. In total there are about 3 billion of these building blocks in the DNA of a bull or a cow. 99.8% of the DNA is identical between cattle within a breed, so only 0.2% is different and this part determines the genetic differences we see in field data and in breeding values. This means that there are around 6 million base pairs (0.2%) that cause the difference between bulls and cows. These base pairs that cause these differences are called SNP's (snips). By the way, also between humans only 0.2% of the DNA determines the genetic differences we see between each other.

Genomics is the method where the information from the traditional breeding value estimation is combined with the DNA information. From a large group of animals with very reliable proofs, the reference population, the DNA is analyzed. In this analysis, about 45,000 SNP's are mapped. With powerful statistical tools the relation between these 45,000 SNP's and each breeding value is estimated. i.e. the relation between each SNP and the proof for %Fat is estimated and the relation between each SNP and Udder Depth proof and so forth. So for each of the 45,000 SNP's the relation with each breeding value is estimated. These relations are called SNP effects. Some SNP's have a very strong effect on percentage fat but have very little or no effect on other traits. When all these little SNP effects per trait are summed, the Direct Genomic Breeding Value can be calculated.

It is assumed that the SNP effects that are estimated in the reference population can be extrapolated to other animals in the same breed. This opens the door to test the DNA of very young animals and calculate genomic proofs, using these SNP effects. How good is this assumption? At the moment the reliability of genomic production and type proofs is around 70%, so the assumption is quite good... but not perfect. However compared to the reliabilities of parent averages, it is a giant step forward. We have worked for many years with parent averages that have reliabilities of 35%. A reliability of 70% can be compared to the field data of 35 daughters of a sire. The chance is much higher that the current genomic proof of a young bull will end up closer to his final daughter based proof than his traditional parent average.

 Although there are still many hurdles to take in the development of genomics, Alta strongly believes in the power of this technology. All of our young bulls are genomically tested upon release.

Today, dairy producers have a better set of options to choose from. On the one hand, they can choose to use high reliability daughter proven sires with reliabilities above 90% and generally more moderate genetic merit. Or they can they can choose to use G-Star and FutureStar bulls and access the highest GPTI or GNM$ bulls available with a reliability of 70% - but combined as a group of 5 can reach over 90% reliability for the average.

Article by Gerbrand van Burgsteden, Product Development