Performance Testing

________________________________________________

(Compiled from the BREEDPLAN MANUAL with slight adaptations - C P Massmann)

What is the difference between indexes and EBV's (Estimated Breeding Values)? 

An index is used on a within-herd contemporary group basis only and it is not a genetic measure.  Indexes can only be used for within-herd comparison in a particular season. EBVs consider six major areas that are not accounted for in indexes:

1.  EBVs are genetic measures.
2.  EBVs allow comparison of animals born in different seasons and years.
3.  EBVs allow for the fact that some sires are used on better cows.
4.  EBVs take account of genetic change in the herd over time.
5.  EBVs also incorporate genetic information from all relatives and other connected traits.
6.  EBVs take account of the amount of performance records available for each animal.

Selection on EBVs therefore gives much quicker progress than indexes.

How much information in an EBV comes from an animal’s pedigree and how much from the individual’s own performance?

The information used to calculate an EBV can come from several sources, the animal's own performance, ancestor records, paternal and maternal half-brothers and sisters, and progeny records.  Each source is appropriately weighted according to relationships (all through the sire and dam), between the relatives and the animal itself.  Once an animal has many progeny, less emphasis is placed on pedigree information and the animal's own performance and the performance of his progeny become the major determinants of his EBV.  With highly heritable traits the animal's own records are more important, while with lowly heritable traits (e.g. fertility) records on relatives and related traits become more important.

Does BREEDPLAN make allowance for seasonal variation and differences in nutritional levels for the herds involved?

Each animal is initially compared only with animals of the same sex born within a limited calving spread and reared under the same conditions.  It is then compared to other groups.  A calf which is 25kg heavier as a yearling than its group average of 350kg in a poor season, will get the same credit as another calf that is 25kg heavier than a group average of 400kg in a good season.

The variation in nutritional levels is accounted for by the use of link or marker sires which have progeny in at least two herds.  The link sires in each herd are used as the reference point for that herd and all animals are ranked in relation to these link sires.  The fact that the progeny of a link sire weigh an average of 240kgs at weaning in one herd and only 180kgs in another herd does not disadvantage the herd with lower nutritional levels.  In fact herds from less favourable environments have a better opportunity to demonstrate their genetic potential through BREEDPLAN.

What is the accuracy of an EBV?

The Accuracy or reliability of an EBV for a particular animal is determined by the amount of information available about that animal, its relations, and the animals it is compared with. The more information that is known, in particular the greater number of progeny that are analysed, the higher the heritability of the trait and the more animals in the comparison, the more reliable are the EBVs.

EBVs with low Accuracy may change quite markedly (up or down) with the addition of more information, whereas EBVs with high Accuracy are unlikely to change much. However, it is important to remember that there is an equal chance of EBVs increasing or decreasing as more information is added.  Accuracy then is a measure of risk.  If you are only selecting one sire to join with heifers accuracy would be more important than say buying three young bulls for a group of cows.

Whilst performance records from relatives such as half sibs (brothers and sisters) do increase the accuracy, substantial increases in accuracy are only realised when progeny records are added.

If the only information available is the bull's own 400 day weight (a trait with a heritability of 0.3), the accuracy of his EBV for that trait will be 55%.  If 10 progeny records are added to the analysis the accuracy of his EBV will increase to 74%, compared to only 64% if records of 24 half sibs are added.  See table for the effect of additional performance records on the accuracy of an EBV for a trait with heritability of 0.3.

Performance is measured on: Accuracy (%)
   
Individual 55
Individual + 10 PHS* + 2 MHS** 61
   
10 progeny 67
32 progeny 85
   
Individual + 10 progeny 74
   
Individual + 45 progeny 90

*        PHS is paternal half sibs or other calves by same sire
**       MHS is maternal half sibs or other calves by same dam

(The GROUP BREEDPLAN or national analysis accuracies will be higher than in the table because the analysis also uses information from other correlated traits).

How much attention must I pay to the accuracy of a trait?

As a general rule, animals should be compared on EBVs regardless of accuracy, however where two animals have similar EBVs the one with the higher accuracy could be the better choice, assuming other factors are equal.

When you need to choose a single sire that will perform to expectation (e.g. low birth weight sire to join to heifers), choose a sire with both the desired EBVs and high accuracy for the traits of importance as there is less chance that the EBV will change as more information becomes available.

However, if you can afford to sample a  number of sires then you should consider choosing sires with the best EBVs for the traits of interest with little regard to accuracy.  Even though the individual EBVs of these bulls could change, their average EBV will not.  You can expect that EBVs for 1/3 of the bulls will in fact be higher than first estimated thus you have a greater chance of finding a real herd improver for the target trait.  Of the other bulls, EBVs will remain steady for 1/3 and decrease for the other 1/3.

When buying young bulls e.g. 18-24 months, their accuracies for a particular trait will generally be quite similar.  In these circumstances even the single bull buyer can concentrate on the EBV rather than accuracy.

EBVs for an animal have changed considerably since the last analysis

Please bear in mind that the accuracy of EBVs improves as more information becomes available in the calculation of the EBV.  Here are only a few answers :

    ■  EBVs for calves will change when more performance is submitted for them.  If little is known about the parents of these calves, the EBVs of the parents will also change to
         some extent.

    ■  EBVs for sires without any, or many progeny observed in prior analyses will change to some degree each time more performance recorded progeny are recorded for him. 
        The same applies for dams.

    ■  The Milk EBV for sires will change to some degree each time more 200-day weights a are recorded for his daughter’s calves.

    ■  Significant changes to the performance records of a calf may lead to a marked change in the EBVs of close relatives, depending on the amount of information already on
        known about the relatives.

    ■  Recording errors in new herds can have a large effect on some groups of animals, e.g. lack of management group recording for commercial cows.

High or low mature cow weight EBV’s?

Mature Cow Weight EBVs are calculated from weights taken on the cow when her calf’s weaning weight is being measured.  In this way the analysis looks at the weight of the cow and the weight of the calf she produced.  The Mature Weight EBV is standardised to indicate differences between cows at 5 years of age.

Selection for growth rate is likely to increase the mature size and therefore maintenance feed requirements.  Mature size EBVs would allow breeders to identify those animals with high growth rate and moderate mature size.  The Mature Cow Weight EBV would be an indicator of:

    ■  cow feed requirements   -   until feed efficiency EBVs are available, many breeders may try to reduce the mature size of their breeding cows while maximizing the growth
         to say 400-days.  Yearling breeders for example may seek stock which “bend the growth curve”, with high yearling weight, but moderate mature cow size.

    ■  Mature steer size   -    breeders of heavy steers may look for large mature weight EBVs – particularly if looking at long term feeding programs.

It can be expected that a cow with a higher Mwt EBV may require more feed to maintain her than a cow with a lower Mwt EBV, she is also more likely to have a heavier calf at weaning. A rough guide is that Mwt EBV should be less than the same animal’s 600-day Wt EBV. For example a cow that has a lower Mwt EBV than its 600d EBV is able to pass on its growth to its progeny while maintaining a lower mature size.

Two of my bulls ,weighed on the same day in the same group, have the same 400 day weights but different EBVs…?

Two bulls (A and B) both have a 400-day adjusted weight of 380 kg compared to the average for the group of 350 kg. If there were no information on the genetic merit of the parents or other relatives, both animals would have an EBV of +9kg for 400-day weight.

[ (380kg-350kg)* 0.30 heritability = +9kg]

However, if there was performance information on the parents, the parents' EBVs will influence those of these two bulls.  The bulls’ information will also affect the parents' EBVs but the higher the accuracy of the parental EBVs the less this effect will be.  In the following example, high accuracy parent EBVs are assumed.  e.g. parents of bull A have EBVs of +12 and +4 for 400-day Weight and the parents of bull B have EBVs of +4 and +2.  In other words the parents of Bull A are genetically superior to the parents of Bull B.  Given this information on parents, the EBVs for 400-day Weight for the two animals will be adjusted to something like:

  Initial Calf EBV Parents Average EBV Calf EBV adjusted for parents
         
Bull A +9 +8 12 +11
      4  
Bull B +9 +3 4 +7
      2  

Although the two animals had the same individual performance, the superior genetic background of bull A indicates that his progeny will probably be a little superior to the progeny of bull B.  Because both parents are known, the EBV for the calf is made up of the midparent EBV and a portion of the difference between the calf's measurement and the group average.  This portion is worth only half as much if the parents are unknown.

“I weigh all my animals on 400 and 600 days . Why do my 400-day EBVs change after the 600-day EBVs have been released?”

This is called the effect of correlated traits. Two calves have the same EBV for 400-day Weight after yearling weights are analysed.  The EBVs for 600-day Weight have been predicted from the 400 days weights and the performance of parents.

 

400-Day Weight EBV

600-Day Weight EBV

Calf A +9 (+15)
Calf B +9 (+15)

When the calves are weighed at 600-days Calf A has an adjusted weight of 520 kg and Calf B an adjusted weight of 490 kg compared to an average for the group of 460 kg.  With the addition of the 600-day weight information BREEDPLAN recalculates the EBVs for 600-day Weight and also the EBVs for 400-day weight because of the genetic correlation between 400 and 600-day weights.  The superior performance of Calf A at 600 days of age gives it a high 600-day EBV and also causes a slight increase in its EBV for 400-day Weight.  The reverse effect occurs for Animal B because of his poorer performance at 600 days.

After 600-day weights are included in the analysis the EBVs will be adjusted to something like:

  400-Day Weight EBV 600-Day Weight Kg 600- Day Weight EBV
Calf A +11 520 +20
Calf B +9 190 +10

Note:  When actual measurements are available, the correlated traits have a relatively small effect on EBVs

“Tell me more about the milk EBV”

The Breedplan 200-day Milk EBV is the best estimate of an animal's milk-production ability.  In the case of sires, this of course refers to the milking ability of their daughters and may therefore take some time before a sire actually has daughters with calves performance recorded at weaning. 

Until a sire has daughters in production, his EBV for 200-day milk is determined from his relatives' EBVs.  The accuracy of a milk EBV that is based strictly upon pedigree information is low and the possible change in EBV from one evaluation to the next, as daughters are brought into production, could be quite large. The 200-day Milk EBV is also the EBV which is most influenced by poor recording of performance data e.g. neglecting management groups.

The question might arise why Simmentalers should be selected on their milk BTWs since the breed is noted for milk.  The answer is that, as in any other breed, some Simmentaler bulls have more ability to sire daughters that produce more milk than other Simmentalers. 

The 200-day Milk EBV is also the EBV which is most influenced by poor recording of performance data e.g. neglecting management groups.

The optimum level of milk production potential among beef cows is dependent upon the production system and environment in which the cows are run. Selection for increased milk production may be warranted when cows are run under good nutritional conditions and calves are sold as weaners. However, some environments may not support high milking cows.

How large should my herd be to participate?

There is no actual minimum size herd but to get effective results from your own records, you would need about 10 cows calving in a 12 week period. Small herds can increase the accuracy and effectiveness of their results by keeping their calving period as short as possible and by keeping the number of management groups to a minimum. For example, do not castrate bull calves until after the first weighing has been recorded, and do not split the mob into sub-groups unless entirely necessary.

Where only one sire is used in a herd in any year, EBVs cannot be calculated for the sire, but will be calculated for the cows and calves.  Small herds can, however, contribute valuable information on AI sires to the GROUP BREEDPLAN analysis for their breed.

Some guidelines that breeders may follow to obtain effective results:

The basic mechanism by which BREEDPLAN works is to compare the performance of animals to other animals in the same group.  Where there is only one animal in a group there is nothing that it can be compared with and therefore its performance cannot be used.  Calves are only analysed in the same group if they :

    ■  were bred in the same herd,
    ■  are of the same sex,
    ■  were born within 45 days (for birth and 200-day weight) or 60 days (for 400 and 600-day weight) of each other,
    ■  have been run under the same conditions,
    ■  and have been weighed on the same day.

An analysis group must therefore have a minimum of two animals that meet these criteria in order to be used in the BREEDPLAN analysis.  There are a number of strategies that breeders can use to ensure that the performance of calves will be included in effective analysis groups.

1.  Restricted calving periods:  as calves are only included in the same analysis group if they are born within 45 or 60 days of one another, it is essential that small herds have
     as short a calving period that is practical, - 6 to 8 weeks is ideal.
2. 
Run all calves under the same management conditions :  where possible all calves should be run under the same conditions and weighed on the same day.  If calves are
     to be split into different groups it is useful to weigh the whole group before it is split.  For example, it is possible to take 200-day weights anywhere between 80 and 300
     days of age, therefore you can weigh all male calves as a group before a portion of them are castrated.
3. 
Inclusion of commercial animals : If you have a commercial herd, then consider recording these as well as your “stud” animals.  This allows a greater number of animals to
     be included in the same analysis group.
4. 
Use more than one sire :  Another important factor to consider is that a herd should use more than one sire in any joining program.  BREEDPLAN requires at least 2 sires to
     be represented in an analysis group if the performance of the progeny is going to contribute to the calculation of EBVs for their sire.  Where AI programs are used they
     should be timed so that AI sired calves are born at the same time as calves sired by natural joining.
5. 
Genetic Linkage :  As with all BREEDPLAN recording herds, you need to use some sires that have performance records in other herds.  This is called genetic linkage and
     is necessary to allow benchmarking between herds to calculate EBVs.

In many cases, the advantages of recording a small herd in BREEDPLAN is to gauge where the animals within the small herd relate to the rest of the breed.

 

    © 2003 - 2017 The Simbra Cattle Breeders' Society of Southern Africa
    Website Terms & Conditions

 
    Home Page
    The Simbra Breed
  Genetic & Phenotypic Trends   Breedplan Recording System
    Publications
  ______________________________________________________________________     Correct Recording ______________________________________________________________________
    The Society
  Why Simbra?
    Purchase of Animals
  Frequently Asked Questions   Journals
______________________________________________________________________   History ______________________________________________________________________   Birth Weight & Calving Ease   Bulletins
  Objectives   Standard of Excellence   Selection Aid     Pamphlets
  Mission & Vision   Simbra Stats   Establishment of Herd
    Shows & Judges
  Articles
  Breed Recognition   Simbra Brand Mark   Simbra Certificate ______________________________________________________________________  
  Values & Principles   Simbradex   Selection of Bull   Policy & Training
    Links
  2017 / 2018 Council   Polled Animals   Quality Control   Judges, Examiners & Inspectors ______________________________________________________________________
  Office Staff   Fertility     Basic Principles of Judging   Breeders
  Technical Advisors   Published Sires & Trait Leaders
    Performance Testing
  Show Results   General
  Clubs   Star Cows ______________________________________________________________________   Show Classes  
  Register as Breeder   Percentile Table   Information   Show Catalogues
    Calendar