Careful versus careless use of genomic evaluations by dairy producers
Genomic evaluations of young dairy bulls have greatly altered the landscape for artificial insemination (A.I.) bull usage in herds. However, it seems some dairy producers, as well as many journalists who write about genomic evaluations and marketing personnel of A.I. companies, tend to infer the genomic evaluation of a young bull (or heifer) is the “final answer” (or very nearly so) on genetic worth of that individual. In reality, that isn’t the case. For a young bull, the “final answer” for his genetic evaluation isn’t achieved until he has hundreds, if not thousands, of daughters spread across many herds with lactational performance.
Just how much do the genomic evaluations for individual bulls that are highly-ranked change with time? The answer is a lot! I’ve provided the genetic evaluations for the top 10 genomic-only young Holstein bulls that were ranked on the Net Merit selection index in August 2013 (readily obtained on-line from Holstein World) along with the subsequent genetic evaluations of the same bulls 3.75 years later in April 2017 that are based on daughter performance. The #1 bull in August 2013 had genomic evaluations of $683 Net Merit, +654 lb milk, +40 lb fat, and +28 lb protein. Later, in April 2017, when the same bull had 1,271 daughters contributing to his proof, his genetic evaluations plummeted to $277 Net Merit, -955 lb milk, -24 lb fat, and -19 lb protein. Yes, a loss of 1609 lb of milk from a genomic-only to a daughter-based evaluation!
Notably, two of the top 10 genomic bulls in August 2013 ended up becoming giants among proven Holstein bulls 3.75 years later in April 2017 – Seagull-Bay Supersire-ET and Co-op Robust Cabriolet-ET. These two bulls happen to share the same sire (Robust). Supersire was ranked 4th among the August 2013 genomic-only bulls at $670 Net Merit, but in April 2017 increased to $866 Net Merit. Cabriolet was 8th among genomic-only bulls at $646, but in April 2017 increased to $875 Net Merit. Only one additional bull among the top 10 in August 2013 ended up with a higher Net Merit based on daughter performance than on his genomic-only evaluation. All of the other 7 bulls among the top 10 in August 2013 dropped for Net Merit (as well as for all the production traits), and some of them dropped dramatically.
So, what does this tell us? Genomic-only evaluations for individual bulls, especially those ranked near the top of a breed, often change markedly when their actual daughter performance becomes available. Consensus in the industry is the most highly-ranked young bulls at any point in time are more likely to be over-evaluated than those ranked more lowly. This may be contrary to what many have been led to believe. Genomic-only evaluations have fairly good stability for groups of bulls within designated levels (stratifications) for a trait or index, but the genomic-only evaluations for the most highly-ranked individual bulls at any point in time are prone to substantial change.
Therefore, dairy producers deciding to use genomic-only A.I. bulls must make use of a fairly large group of young bulls at all times to spread the risk of change across the group of bulls instead of cherry-picking a small number of highly-ranked individual young bulls. The marketing of individual highly-ranked bulls based only on genomics in the same way that proven bulls (that have a large number of daughters with performance information) are marketed is a questionable practice that isn’t the norm in many other countries. Also, the group of genomic-only young bulls used in a herd needs to be changed periodically within a year. New and updated official genomic evaluations are released at 4-month intervals. If the sire of highly-ranked genomic-only bulls has a substantial drop in his genetic evaluation for a trait or index based on daughter performance, then all his sons will likely drop substantially for that trait or index, too.
Lastly, some dairy producers also seem to believe the genomic evaluations of young heifers are the “final answer” on their genetic worth and, therefore, genomic evaluations are highly reliable to use for selection among heifers. If the sires of heifers are genomic-only bulls, then their daughters may have large changes in their genomic evaluations when their sires eventually have many daughters with performance information. Again, genomic evaluations are updated every 4 months, and they may change greatly from one evaluation to the next for an individual young bull or heifer. Please carefully, rather than carelessly, use genetic evaluations that are based only on genomic predictions.