Calculation methods

BLUP

BLUP, Best Linear Unbiased Prediction method, enables the simultaneous estimation of environmental and genetic influences. As the environmental effects are distinguished from the genetic merit, the environmental effects can be eliminated from breeding value predictions. For example herd's feeding level does not affect the animal's protein production index. The systematic effects which are included in the model besides herd effect are e.g. lactation number, stage of lactation and season. In calculation of each index the environmental factors affecting the traits in question are taken into account.

Animal model

Animal model method includes information of all relatives of the animal. All various relationship data is weighted by their importance. For bulls the most important source of information is the data of bulls' own daughters although all the information of relatives is utilized. Taking into account the information of all known relatives increases especially the accuracy of cow indices. Animal model takes into account also the effect of bull dam selection in bull evaluations. At the moment animal model is used in index calculations for days open, live weight and conformation traits.

Test-day model

From April 2000 onwards breeding values for production traits (i.e. milk, protein and fat production, protein and fat content) and somatic cell count are estimated by using test-day model. In test-day animal model breeding values are estimated directly based on monthly test-day measurements made at farms. Compared to traditional lactational 305 d model, test-day model enables better description of environmental factors that affect the actual test-day records such as stage of gestation, calving age, calving month and especially seasonal variation within herds.

With the test-day model a cow gets the first prediction already after her first test-day record. Earlier breeding value estimates were calculated after a cow had at least 4 test-day records from which a 305 d record was constructed assuming a normal lactation curve. Now there is no need to compile individual test-day yields or extend incomplete lactation records.

Finnish random regression test-day model includes individual test-day records from all lactations from 1988 onwards. First lactation and later lactations are considered as genetically different traits; test-day model estimates genetic lactation curves which describe the breeding values for each animal for first and later lactations separately. The production traits are evaluated simultaneously, i.e. all genetic and phenotypic associations between milk, protein and fat yield at different stages of lactation are taken into account to increase the prediction accuracy especially for protein and fat yield.

The reliability of protein production index calculated with test day model describes the accuracy of the evaluation. The reliability of the index is affected by the heritability of the trait and the amount of information utilized in evaluation. Reliability of the evaluation is improved most by increasing the amount of daughters' production information in evaluation. If the bull has not yet any daughters in production evaluation the reliability of the index is appr. 40 %. If the number of daughters is over 200, and they have several test day observations, the reliability is well over 95 %.

Sire model

When the trait has a low heritability the emphasis is focused mainly to progeny testing and the selection is mainly carried out among sires. The evaluation is calculated with sire model which utilizes the relationships between bulls. Sire model is used in calculations of indices for milkability, leakage, temperament, fertility treatments, mastitis treatments and other treatments. Because these traits are not object for corresponding selection as milk traits, the sire model gives relatively good accuracy in index calculations.

Non-return rate, calving ease and calf mortality indices are calculated using sire-grandsire model BLUP. Calving difficulty and mortality indices are published for bulls as a sire and maternal grandsire of the calf.

Base and standardization of indices

The comparison basis for bull indices is the average genetic level and standard deviations of bulls born from seven to nine years earlier. All bulls of the same breed born during these years form the comparison group with a mean index value of 100 and a standard deviation of 10. All indices of animals of the same breed are standardized to this mean. So the indices are comparable for all cows and bulls of the same breed and of different generation groups. Genetic reference base is annually rolling forward by one year. Animals' genetic progress is considered through the change in genetic reference level.

Additional information through the new evaluations improves the accuracy and precision of the indices. The indices of bulls with a high evaluation accuracy do not show a significant change between the evaluations during the same year but due to a change in reference generation the indices can change somewhat when the evaluation year changes.

Total merit index for bulls

Bulls' total merit index (TMI) is the most important criteria for bull selection. The main goal of bulls TMI is the increased dry-matter production and longevity.

Because longevity and durability are more and more important in milk production the weights of udder conformation and health are higher than earlier. Besides fertility and udder health  also udder and feet and legs indices are included in TMI. The TMI of Holstein-Friesian bulls includes also calving index (maternal grand sire).

TMI of Ayrshire bulls: 0,9 * production + 0,3 * udder health + 0,4 * udder conformation + 0,3 * fertility + 0,1 * feet and legs

TMI of Holstein-Friesian bulls: 1,0 * production + 0,5 * udder health + 0,3 * udder conformation + 0,3 * fertility + 0,1 * feet and legs + 0,1 * calving index (mgs)

TMI of Finncattle bulls: 0,9 * production + 0,3 * udder health + 0,4 * udder conformation + 0,3 * fertility

Total merit index for cows

Cow index is a predicted economic breeding value for cows. It is used in dam selection and for within herd selection of animals to produce replacement cows. The cows of different herds and of different age groups can be compared due to animal model calculations. Profitable dairy cow is a combination of good production, conformation, health and functional traits.

TMI of Ayrshire cows: 0,9 * production + 0,5 * udder health + 0,5 * udder conformation + 0,1 * feet and legs

TMI of Holstein-Friesian cows: 0,9 * production + 0,7 * udder health + 0,4 * udder conformation + 0,1 * feet and legs

TMI of Finncattle cows: 1,1 * production + 0,3 * udder health + 0,5 * udder conformation