Using performance-testing programs in piggeries

Understanding the effects of variation on improving herd genetics can help you to use your on-farm, pig performance-testing programs to make improvements to your selection and breeding program.

If you are serious about genetic improvement, the ideal option is to use information from best linear unbiased prediction (BLUP) and National Pig Improvement Program (NPIP).

A pig herd with the following characteristics is well placed to carry out a genetic improvement program:

  • herd size – 75 sows, 4 boars
  • replacements per year – 40 gilts and 4 boars
  • performance tests – every 3 weeks
  • pigs available per batch – 9 litters with 8 pigs (36 gilts and 36 boars)
  • number selected per batch – 3 gilts (per batch) and 1 boar (4 times per year).

This page includes a detailed description of the topics:

  • general principles of performance testing variations
  • ideal performance-testing conditions
  • possible genetic improvement
  • variations
  • reduced degrees of choice
  • actual performance versus pen margin
  • mixing sexes
  • mixing litters
  • homebred or buy in.

General principles of performance testing variations

  • Regardless of where you get the boars, failure to performance test gilts reduces genetic gain by about 45%.
  • The more matings made to boars from non-testing herds, the more your own efforts at genetic improvement will suffer.
  • Using artificial insemination (AI) boars with above-average estimated breeding values considerably enhances your rate of improvement. Several different AI boars should be used to limit inbreeding.
  • Adopting several reduced-effort variations can significantly affect the genetic improvement of your pig herd.

Ideal performance-testing conditions

An ideal genetic improvement program will have:

  • at least 10 pigs per pen
  • pen mates of the same sex
  • pen mates born in the same week
  • pen mates from several litters by different sires
  • fewer than 15kg (live weight) of difference between the heaviest and lightest pen mate
  • pigs in the same pen during the 50–90kg growth phase
  • breeder selection based on margins above or below the average performance of the pen
  • degrees of choice of at least 20 males and 10 females tested for each boar and gilt selected
  • low competition for food and high health status.

Possible genetic improvement

Given these ideal testing conditions, genetic theory can predict improvement. Selection is a process of replacing boars and sows with their best-tested offspring. A single generation of selection, taking about 2 years, will increase the value of each pig produced in the herd in all future years. Background knowledge of costs and returns of pig production suggests that, on a value of $150 for a baconer, this improvement is worth $12. Additional generations of selection will further increase the value of each future pig.


Reduced degrees of choice

The degree of choice achieved can be less than the ideal of 1–in–20 males and 1–in–10 females tested. Sometimes, the best performing gilts are unsuitable for breeding due to unsoundness or anoestrous. Halving the degree of choice to 1–in–10 males and 1–in–5 females reduces the genetic gain to 84% of that possible for the ideal program. In practice, a degree of choice of 1–in–20 for boars and 1–in–5 for gilts can be readily achieved. This gives 91% of the possible genetic gain.

Actual performance versus pen margin

Pigs should be assessed as future breeders not according to actual performance but according to pen margin. This is the difference between their actual performance and the average performance of the pen in which they grew. This simple calculation removes the inaccuracies caused by growth environments, which vary between pens. If we assess pigs on their actual performance rather than pen margins, we may reduce genetic gain significantly.

For example, if all pigs were selected on their own performance, genetic gain could be reduced to 85% of the ideal. Selecting gilts this way would reduce genetic gain to 95%.

Mixing sexes

Pens containing pigs of both sexes can cause difficulties for a performance-testing program. As males and females perform differently, combining them distorts the pen average and the margins above or below the average. It is better to use single-sex pens if enough pigs are available, as this will give the desired degree of choice.

In an ideal herd (as described above), 20 boars would be tested for each 1 selected and 30 gilts would be tested for every 3 selected (1 in 10). If mixed-sex pens must be used, pigs of the same sex in the same pen should be treated as a separate group (if each group contains at least 5 pigs). Sometimes pigs are in such short supply that groups of less than 5 have to be used. The test results of these pigs should be combined with similar small groups of the same sex in other pens. This should be done only if the average growth rates of these groups differ by less than 10%.

Mixing litters

When performance testing, it is good practice to mix pigs from different litters in the same pen. This increases genetic variation and, therefore, the chance of selecting breeders with true genetic superiority. If all pigs within each pen had the same sire but were from different dams, the efficiency would be reduced to 83% of the ideal. However, if all had the same sire and dam (e.g. litter mates), the program's efficiency could be reduced to as little as 56% of the ideal.

Homebred or buy in

Sometimes farmers do not wish to performance test. Gilts are often chosen purely on physical soundness and boars are introduced from herds that may or may not have a performance-testing program. Alternatively, some boar power might be drawn from an AI centre.

Several combinations of home selected and introduced breeding stock were compared (based on 1996–97 figures, i.e. from farm test gilt figures, Wacol Pig Test Station-approved boars). Table 1 shows the genetic gain percentages expected from these combinations. The level of gain expected from our ideal herd, in which all replacement breeders are home tested and selected, is 100%.

The 'untested herd' referred to in Table 1 starts with the same genetic level but is not performance tested and is unlikely to improve. The AI sires are from performance testing herds.