Further discussion on May sire summaries
Holstein World guest writer David Selner answers questions that many of you asked following the release of the May Sire Summaries. If you””d like to read more about this, or have questions that are yet unanswered, email kknutsen@dairybusiness.com and we””ll keep this blogging session open as long as there is interest.
Genetic Evaluations – The Rest of the Story
By David Selner
As reported in the June Holstein World, adjustments made by the USDA and the Holstein Association could have created real confusion in the ranking of sires. However this writer and others did not see any major changes or problems in the sire listings.
But upon hearing from breeders and others, a real confusion over cow indexes has been noted by many in the industry. This difference is caused by the fact that a cow index focuses solely on one individual, where a sire evaluation encompasses many individual cow evaluations. The differences in cow indices were not noticed as much because they are averaged out over many daughters in a sire evaluation. So unless a sire was used to create bias across one group of lower animals versus another with higher values, you would not notice the difference.
What happened to the cow indices on those cows that dropped significantly? The USDA Animal Improvement Laboratory is always doing research to fine-tune and evaluate the accuracy and predictability of the genetic information they publish. This has lead to numerous improvements to genetic evaluations over the last 30 years. Because of a growing awareness of a certain segment of the industry utilizing crossbreeding and the continual effort to provide more accurate evaluations for the genetics industry, the primary focus of USDA scientists lately has been in those two areas.
In the May evaluations a major shift to the use of an all-breed animal model instead of a separate animal model for each breed, was implemented. This allowed the inclusion of many more crossbred animal records and did increase the accuracy of evaluation over previous techniques. Because the predominant breed is Holstein, the new all-breed base is made up of mostly Holstein data, so you did not see any real impact on Holstein sires. Several of the other breeds did notice changes in their sire rankings because of this new base and the additional data being used.
So the change in Holstein cow indices is not due primarily to the model change. However if you were a herd that is actively utilizing crossbreeding or purchasing animals from herds like that, you could have been affected by a model change. The reason is that certain groupings of contemporaries could be altered because of the additional daughter data now being utilized. So who certain cows are compared with could have changed. This is probably not a large segment of the cows whose indices changed dramatically.
The USDA also changed the genetic trends for DPR, calving ease and stillbirths. This might have changed the values of these traits some but would not change the rankings very much.
Another less publicized enhancement applied this time was a revision to the herd variance adjustment used in the animal model. Animals that came with extremely high yield deviations or from herds with unusually high or low herd variances were affected by the change. These are the type of records that offer the least accuracy or predictability of genetic superiority in the current animal model. So adjustments were made to allow these records to be utilized but with different amounts of emphasis because the true genetic effect of deviation is the that which can be transmitted and not the absolute amount of the deviation. There are numerous herds that, by their level of management, experience normally high variation or low variation. You can go into one herd and find animals that are +10,000 pounds of milk over herdmates while other herds you never see a cow over +5000. So the absolute size of the deviation within any given herd can be mostly attributed to management and not transmittable genetics. Therefore these records are now adjusted to a more average variance found in herds to eliminate the upward bias some records could have on the entire system. This is a continuation of improvements that were first starting to be implemented over 15 years ago.
Now those adjustments would have a fairly major impact on the new cow index for that individual animal. So if you had a cow under those conditions her index obviously dropped and you wondered why. This small adjustment affected you but in reality did not impact a larger number of breeders to create a universal concern.
Is this change only temporary or will something alter it for August? I believe that this adjustment is a permanent change to the system and will remain in effect for the future. Remember the accuracy of the genetic values to predict the genetic value of an offspring is what the ultimate goal. These changes help improve that accuracy.
What is going to happen next and how will I be affected? That is the question that is being asked by many of these breeders affected by this variance change. Now I do not know what improvements will happen and when. However the USDA has already given an indication that it is moving towards a test day model. These same scientists are continuously looking at the accuracy of international genetic evaluations and their predictability globally. If either of these areas create a change which of my cows is going to be affected the most? Again this is nearly impossible to predict until you see all of the things that will be impacted by such a model change.
So the big question some breeders pose is how can I stay a supplier of elite genetics if I do not know how to predict who is going to be affected? The simple answer can be expressed in three key areas. First, you should just utilize the most accurate of current genetic rankings to select for the future. Next, provide an environment where every animal gets an equal chance to demonstrate her genetic superiority. Finally, realize that future genetic rankings are going to include many different traits and not just the old standbys of milk, component test and type.
No matter what system we utilize in the future, the goal will be to identify those cows and bulls that perform most profitably in the management systems and cost structures being utilized in this country.