|
May 2007 Genetic Evaluation Changes
A number of improvements to genetic evaluation procedures are being implemented with the release of the May 2007 genetic evaluations. In fact, the level of revisions is so significant, the May evaluations should be viewed similar to new evaluations following a base change. We probably have not seen this level of modification since Animal Model evaluations were introduced in the late 1980’s.
You can be assured that a high level of testing has occurred validating the accuracy of these new evaluation procedures. Evaluation model changes implemented by USDA and the Holstein Association have met the Interbull standards for genetic trend testing and have improved the consistency of results when genetic values are compared across countries. In addition, test evaluations using data available from the February evaluations, have been available for industry review since early April. Based on these results and industry review, the decision was made to move forward with these modifications in May.
This document explains the evaluation changes and their impact on sires marketed by Select Sires.
USDA All-Breed Animal Model Evaluations
Records from cows of all breeds, including crossbreds, are now combined and analyzed together in one animal model. Traits evaluated with this new model include milk and component yields, productive life (PL), somatic cell score (SCS) and daughter pregnancy rate (DPR). Evaluations are calculated initially on an all-breed base and then are converted to traditional within-breed genetic bases for publication. For instance, Jersey cows and bulls and crossbred cows with Jersey sires will have evaluations that deviate from a base group of Jersey cows. A different base will be used for Holstein bulls and cows, Brown Swiss bulls and cows, etc.
Using an all-breed approach to genetic evaluation is beneficial for several reasons. The evaluations allow for more direct breed comparisons and the adjustment of heterosis for crossbred animals. For breeds other than Holstein, the previous evaluations included some crossbred animals with no adjustment for heterosis. The new evaluations will more appropriately use records from crossbred cows in the evaluations. An all-breed approach also allows the use of crossbred animals in the evaluations while utilizing all the available information on the parents of these animals. This would be very difficult if these animals were used in evaluations conducted separately for each breed. Including crossbred animals significantly increased the number of cows available for sire evaluation in the non-Holstein breeds. In addition, the all-breed model allows cows of different breeds to serve as herdmates so that more cows are used in making management group comparisons in mixed-breed herds. This leads to more accurate herdmate deviations especially in cases where a breed makes up only a small portion of the herd. Finally the all-breed model allows the evaluation of crossbred animals sired by bulls from newly introduced breeds (e.g. Swedish Red, Montbeliarde, etc.). When enough data accumulates, this new evaluation system will provide valuable information about the performance level of these breeds.
Unknown Parent Groups
The new evaluation model required changes in other parts of the evaluation or provided a good opportunity to make improvements. Grouping of unknown parents is needed in calculating genetic evaluations and is quite an important part of animal model evaluations. In the new model the unknown parent groups were redefined and now are separated by breed, pedigree path, national origin and birth year. The new model allowed these definitions to be more specific and accurate especially for the non-Holstein breeds.
Within-Herd Variance Adjustments
Herds and breeds differ in the amount of variation in their records. An adjustment to correct for this variation difference has been in place for some time but is modified with the all-breed model. These adjustments are individualized for each herd if enough completed lactation records are available to estimate the level of herd variation. Previously, records with a data collection rating (DCR) of 95% or higher were used to estimate this variation. This requirement has been reduced to a DCR of 92% to make more records available to estimate within herd variation. Some herds on AM/PM testing had few or no cows included in the herd variation estimate in previous evaluations and were adjusted to a general breed by region average variance. Under the new model more herds will receive a herd-specific herd variance adjustment. This revision may cause evaluation changes in cows with extreme yield deviations or in herds with unusually high or low variation.
Corrected Parity Adjustment for Cows that Changed Herds
A separate correction to the calculation of parity adjustments for cows that changed herds resulted in about 3% lower genetic trend for all traits except PL. That lower trend frequently resulted in considerable change in PTAs of animals that were born in the 1960s and 1970s and less change in the opposite direction for most animals of current interest.
New Genetic Parameters for DPR
Several genetic parameters were adjusted for DPR so that the evaluations are better at meeting the genetic trend validation tests. The new parameters make the DPR evaluations more consistent with the genetic rankings from other countries and the genetic trends match more closely with other research estimates. These adjustments caused an average change in PTA for DPR of -0.3 in Active AI Holsteins and Jerseys.
Holstein
All-Breed Model
The impact of the USDA evaluation changes are summarized in Table 1. This table lists the average change in proofs for February 2007 Active AI sires. The use of cross-bred records had little impact on Holstein evaluations. The average change for the production traits, PL and SCS is virtually zero. The expansion of the herd variance adjustments and the corrected parity adjustments for cows that changed herds did impact the genetic rankings especially for cows. The changes to the DPR evaluations caused a general decline in DPR PTA which leads to decreased values for NM$ and TPI.
PTAT Evaluation Changes
A couple of modifications were made to the PTAT evaluations to make them more accurate. First, additional information was assembled to fill in the pedigrees for unscored dams of cows. Previously these dams were simply treated as unknown. An unscored dam may be fairly rare in registered herds but in our progeny test herds where the SET’s are conducted there would be a pretty high frequency where the dams of our PGA daughters have never been scored. Many times the pedigree information exists in the DHI records and with these new evaluations an effort was made to assemble this information and make the pedigree file more complete. Also, modifications to the evaluation software were made to allow more complete processing of the type evaluations.
These modifications produce some pretty big changes in sire evaluations. The average top-400 TPI sire will see a +0.4 change in his PTAT. Second-crop bulls will generally increase more than this.
The new PTAT evaluations greatly improve the stability of proofs when first-crop to second-crop evaluations are compared. The average change for bulls moves closer to zero and about an equal number of bulls go up as go down. Also, the US type evaluations are now more consistent with evaluations from other countries. At this point, these evaluation changes apply only to the PTAT evaluations and will likely be incorporated into the linear trait evaluations in the future.
New Udder Composite
In August 2006, evaluations for rear teat placement were introduced. Research indicates that too wide rear teats or too close rear teats leads to increased SCS. So, rear teat placement is being added to Udder Composite as a trait with an intermediate optimum. Bulls that have STA’s for rear teat placement of +1.00 are considered optimal and those that deviate a lot from +1 will have reduced udder composite values. While rear teat placement is added to the composite, the emphasis on front teat placement has been reduced and the emphasis on udder depth has slightly increased.
New TPI™
The Total Performance Index (TPI) has been revised to place more emphasis on improving reproductive performance and calving ability in the Holstein breed. The new index now includes a small weight on Daughter Stillbirth evaluations and increases the amount of weight on DPR and PL. To make room for the extra selection on fitness traits, emphasis on the production traits, particularly protein yield, has been reduced. There has been no change on the weighting on type traits.
Jersey
All-Breed Model
Table 1 includes the changes of proofs on Active AI sires and shows that we can expect to see a general decrease in PTA’s for Jersey bulls for production traits, PL, DPR and NM$. The decline is due to a more appropriate heterosis adjustment in the evaluations, more known parents for crossbred animals used in the evaluations and simply more animals included in the evaluations. The average Jersey bull will change by -6 for PTA protein and -60 for NM$.
Jersey Udder Index™ (JUI™)
The Jersey Association will begin publishing rankings of bulls for udder conformation using JUI. JUI is based on the weightings of udder traits previously used in their Functional Trait Index. It contains a heavy emphasis on udder depth, rear udder height and rear udder width. For Active AI sires, JUI values range from -3.63 to +8.18 with an average of +2.75. Jersey bulls in Select Sires lineup average +3.94 for JUI.
Brown Swiss
All-Breed Model
The Brown Swiss breed is probably impacted more by the new all-breed model than any other breed. Most first-crop bulls had limited daughter numbers and a surprising number of these daughters had some Holstein in their pedigrees. Table 1 includes the changes expected in the various traits for Brown Swiss. The decline is due to a more appropriate heterosis adjustment in the evaluations, more known parents for crossbred animals used in the evaluations and simply more animals included in the evaluations. The average Brown Swiss bull will change by -9 PTA protein and -89 NM$. The level of impact varies quite a bit for individual bulls and re-ranking of bulls and cows will be noticeable.
Progressive Performance Ranking (PPR)
The Brown Swiss Association is changing the name of its breed index and changing the formula. The new index is called the Progressive Performance Ranking. The new ranking is similar to the old one except that it no longer includes PTAT in the index and puts more weight on PL and DPR. The index relies on udder composite and feet and legs composite for its type contribution.
Body Depth
The Brown Swiss Association has decided it will no longer publish PTA’s for the linear type trait Body Depth as part of the official type evaluations. In the Select Sires materials, a blank line will display in the linear trait block where Body Depth used to be.
Summary
Major revisions are being introduced with the release of the May evaluations. This is a convenient time to remind ourselves that the most valuable information from genetic evaluations is the ranking of animals and the differences between bulls; not the absolute value of the PTA. Try to focus on the ranking of the sire within the breed and not so much on how much he changed from February. The new evaluations are based on more compete data, the evaluations perform better on the genetic trend tests and match more closely with evaluations calculated in other countries. Selection standards, especially for Brown Swiss and Jerseys may need to be adjusted to properly use the new evaluations.
™TPI is a trademark of Holstein Association USA.
™Jersey Udder Index and JUI are trademarks of American Jersey Association.
|
