This press release came from NZAEL (New Zealand Animal Evaluations Ltd) in April:
“The NZAEL has decided to suspend genomic data from BW estimates [Breeding Worth= their equivalent to $Net Merit] until it is better able to provide comparability between estimates of merit of young sires” quoting Bill Montgomerie, head of NZAEL’s dairy evaluation team. Sire proving of the first genomic bulls proved that their real BW was not what had been estimated by genomics. Genomic bull data has now been removed from calculations of the national (NZ) breeding index.
The Kiwi’s had the first operating Genomic estimation system, and the Oceanic AI systems were using this data for young sire selection six years ahead of its release in Europe and North America. However, on their first try, it was found that the average overestimation at “elite” levels was 25-30% depending on breed. They were the ones who recommended to USDA to use a larger number of SNPs to gain on the accuracy—yet the above announcement suggests even that solution is not working for them.
The US and Canadian AI systems are deeply immersed in Genomic theory as the preferred way to select the next generations of young sires for AI. Yet experiences from NZ, and to a lesser extent Europe, are still suggesting dairymen need to be cautious in their selection of young sires on G index rank alone.
Progeny proven sires at high levels of Rel%-- and basing selection on more than just a ranking index—remains the “gold standard” for herd improvement. Population genetics’ methods remain more adept at predicting the behavior of populations than of individual animals.