Mælidagalíkan fyrir íslenska kúastofninn

Many countries utilize random regression (RR) models for genetic evaluation of dairy cows but in Iceland a lactation model is still used. The objective of this study was to investigate the genetics of lactation curves of Icelandic dairy cows and the possible benefits of implementing an RR model for genetic evaluation in the population. 480,495 test day yields of 33,052 cows were used to estimate genetic parameters for milk yield (MY), fat yield (FY), protein yield (PY) and somatic cell score (SCS) in first three lactations with a RR model based on Legendre polynomials. For comparison 305 day lactation yields (LAC1) were calculated from the same data and genetic parameters estimated with a multitrait model. Parameters for MY, FY and PY using the lactation model currently in use (LAC2) were additionally estimated, one trait, three lactations at a time. Breeding values were estimated using the three models with all available data from the years 1995-2016 and for 24 additional versions of that dataset by eliminating data from the last two to eight years, quarter of a year at a time. Correlation between results using the whole dataset and results using the sub-sets was used as a measure of accuracy of estimated breeding values for bulls in progeny testing with artificial insemination and cows in production. Additive genetic variance was highest in the beginning of lactation for all traits in second and third lactation and FY and SCS in first. Permanent environment variance was highest in the beginning of lactation for all traits. Heritability of daily yields and SCS was lowest in early lactation in all lactations and highest in mid or late lactation. Heritability of first lactation 305 day yields was 0.43, 0.39 and 0.41 for MY, FY and PY, respectively, based on the RR model, 0.41, 0.32 and 0.39 based on LAC1 and 0.31, 0.27 and 0.27 based on LAC2. Heritability of SCS in first lactation was 0.23 according to the RR model but 0.15 using the lactation model. Heritability of persistency of lactation MY, FY and PY was 0.14-0.24 in all lactations and genetic correlation to whole lactation SCS -0.08 to -0.13. Considerable difference was found between the models in the age of bulls in progeny testing when accuracy of their estimated breeding values increased fastest. The increase was for the youngest bulls when the RR model was used but oldest when the LAC2 model was used. Predicted genetic progress for PY was fastest by choosing bulls 68, 71 and 83 months old for the RR, LAC1 and LAC2 models, respectively. The predicted progress was more than 11% faster using the RR model than LAC2 and more than 6% faster using LAC1 compared to LAC2. Accuracy of estimated breeding values for cows in production is higher in the period from 3 to 18 months after first calving using the RR model and from 6 to 18 months of age for the LAC1 model, both compared to results of the LAC2 model. Heritability of yields of Icelandic dairy cows has increased and improving the models that are used for evaluation can better isolate the genetic part of yield variation which can improve genetic evaluation. Genetic variation of persistency in the population makes change of the lactation curve possible through selection. Earlier use of records and thus shorter generation interval is the main reason for more predicted genetic progress using the RR model compared to LAC1 and LAC2, and LAC1 compared to LAC2. Application of RR presented here will be beneficial for selection for production traits and SCS of Icelandic cows but genetic evaluation for fertility needs improvement if progress there is not to be reduced.

Author dc:creator

• Jón Hjalti Eiríksson 1991-

Contributors dc:contributor

• Landbúnaðarháskóli Íslands

• Random regression
• Somatic cell count
• Milk yield
• Genetic progress
• Mælidagalíkan
• Frumutala
• Afurðir
• Kynbótamat
• Erfðaframför

Language dc:language.iso

en

Handle dc:identifier.uri

http://hdl.handle.net/1946/29227

OAI identifier oai:identifier

oai:skemman.is:1946/29227