GENOMICS HAS hit the dairy breeding industry by storm.
During the past couple of years, we have gone from almost no application of genomic/genetic marker results in breeding programs to widespread use of genomics.
Genomics provides an opportunity to increase the accuracy of genetic estimates, especially at a young age. This means the ability of breeding programs to make more rapid genetic progress by reducing the generation interval.
However, interpretation of genetic results is not always straightforward in these early days of genomics. Some of the results are difficult to explain.
Despite the large upsides which are well documented – many from simulation studies – some real-world genomic results, identifies areas that still need improvement and where breeders may want to exercise caution in how they apply these results to their breeding programs.
The genetic merit of most of the very high genomic young bulls tends to drift downwards as they add milking daughters. We don’t have a clear understanding of why this happens. This drift may be caused by bias from bull-dams. There may also be statistical challenges of combining data from animals that have genomic information with animals that do not have genomic information.
The good news is that after milking daughters are added, the traditional Total Performance Index (TPI) and the genome-enhanced TPI are relatively close. Despite this downward trend, the rank of the young sires does not appear to change greatly and the genetic merit – even after the downwards drift is still quite high – which is also good news.
Results to date suggest that producers should slightly lower their expectations of high genetic merit from young sires. To avoid disappointment from any one young sire, using a portfolio of several young sires will help minimize risk of any large changes in genetic merit.
Do we need to progeny test now that we have genomics?
Young sires have always been a good genetic option to use in breeding programs – even before genomics. We have been monitoring the performance of young bulls, both with and without genomics, before and after they have had daughters added to their proofs.
One of our interests at ABS is to help determine how extensively we should use young sires as sires of sons (elite breeding bulls).
We looked at 276 young sires tested at ABS over a 15 month period – late 2005 through 2006. As a group, the high genomic young sires were a slightly better predictor of future performance than traditional parent average.
This suggests that both traditional parent average and genomic parent average have merit when selecting young sires, but genomic estimates are a slightly better predictor. Daughter information is still very helpful in finding the outlier bulls.
ABS is using high genomic, young bulls as sires of sons, but we are also using the very highest-proven bulls as sires of sons to help spread risk and to increase diversity.
When genotyping was first made available in 2008, the only option was with the 50k SNP chip from Illumina.
The cost to genotype using the 50k chip back in 2008 was around $US250. The cost to genotype an animal with the 50k chip is now closer to $US150. Additionally, there are now multiple genotyping options.
Illumina now has both a higher density option (800k SNPs) and a lower density option (3k SNPs). Approximate cost in US for genotyping with 800k SNPs is $350/sample and for genotyping with 3k SNPs is $40/sample.
A recent USDA study found that the correlations of breeding values using a 3k subset of the 50k SNPs and the actual 50k SNPs were quite high – ranging from 0.95 to 0.97 (VanRaden, Council on Dairy Cattle Breeding, April 2010). So the 50k chip provides around 96% of the reliability gain compared to using 500k SNPs.
The bottom line is that the cost of genotyping still adds up to a lot of money if a lot of animals in the herd are going to be genotyped. Therefore producers should consider lower cost options first (such as traditional genetic merit estimates) when trying to decide about what animals to genotype.
Genomics is not the only tool available. However, using this tool optimally requires an understanding of the strengths and weaknesses of it as well as the costs. Additional fine-tuning will likely improve the accuracy of the genomic estimates, and further reductions in costs will allow for greater use of genomics for both elite breeding programs, as well as more commercial operations in the future.
Denny Funk is Group Chief Scientist at Genus. This is an extract from his presentation at the recent Herd ’11 conference. He was brought to Australia by ABS Australia.