Estimation of Pedigree Errors in the UK Dairy Population using Microsatellite Markers and the Impact on Selection

Abstract

The proportion of cows in the UK dairy herd whose sires were misidentified was estimated using DNA markers. Genetic marker genotypes were determined on 568 cows (from 168 milk samples and 400 hair samples) and 96 putative sires (from semen samples). The estimated pedigree error rate from the hair samples was 8.8%, and from the milk samples, 13.1%, giving an overall estimate of the error rate of 10%. This level of pedigree errors will have a relatively large impact on the efficiency of progeny testing and the accuracy of cow predicted breeding values. We predict a loss of response to selection of approximately 2 to 3% given this error rate.

Untitled

“…Polymorphism information content ranged from 0.575 (SPS115, 8 alleles) to 0.816 (TGLA227, 12 alleles) with average value 0.713, which confirms a high polymorphism of each analyzed microsatellite. Our results are comparable to those of Holstein cattle found by Heyen et al (1997) in the USA, Visscher et al (2002) in the UK, or Czerneková et al (2006) in the Czech Republic, and higher than those published by Radko et al (2005) in Poland. Generally, the higher the heterozygosity, the higher the genetic variation of the population and its genetic polymorphism, and the more suitable is the marker for individual identification.…”

“…Twelve calves shared one allele with both parents, but it was not possible to determine which one (or both of them) was incorrectly recorded. This result indicates that the interchange of calves on farms (factor (7) according to Christensen et al, 1982) with large-sized herds may play the role as important as the errors in sire identification, or poor recording, which was reminded by Visscher et al (2002). Also Ron et al (1996) stated that the switching of two calves born on the same day was a possible cause of misidentification in large-sized herds.…”

“…If the maternal allele is different from the sire's observed allele, then no common allele will be found in the sire and his progeny, and paternity will be erroneously rejected. Visscher et al (2002) concluded that some of the above-mentioned causes could be prevented by a good recording and verification system at AI. The same authors also pointed out that only factor (7) would give rise to the equal identification error of sires and dams.…”

“…Van Vleck (1970) and Christensen et al (1982) showed that the bias in heritability from half-sib groups was approximately (1-p) 2 if p was the error rate in the progeny of the bulls. Visscher et al (2002) summarized a number of examples of reliabilities and losses in genetic gain as a function of the number of progeny n, h 2 , and error rate p. For example, for an error rate of 10%, heritability of 25% and 50 progeny per sire, the reliability is reduced by 5%, and genetic progress is reduced by 3% (relative to 0% errors). Israel and Weller (2000) presented the same loss in selection response (3-4%) via the …”

“…This expectation is not always fulfilled as several studies showed. The proportion of misidentified progeny varies between 5 and 15% in Denmark (Christensen et al, 1982), 4 and 23% in Germany (Geldermann et al, 1986), 8 and20% in Ireland (Beechinor andKelly, 1987), 12% in The Netherlands (Bovenhuis and Van Arendonk, 1991), 2.9 and 5.2% in Israeli Holstein (Ron et al, 1996), 10% in UK dairy herds (Visscher et al, 2002) or 11.7% in Israeli Holstein again (Weller et al, 2004).…”