|Van Tassell, Curtis - Curt|
Submitted to: Journal of Dairy Science
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 2/17/2009
Publication Date: 7/1/2009
Citation: Wiggans, G.R., Sonstegard, T.S., Van Raden, P.M., Matukumalli, L.K., Schnabel, R.D., Taylor, J.F., Schenkel, F.S., Van Tassell, C.P. 2009. Selection of single-nucleotide polymorphisms and quality of genotypes used in genomic evaluation of dairy cattle in the United States and Canada. Journal of Dairy Science. 92(7):3431-3436.
Interpretive Summary: A total of 40,874 single nucleotide polymorphisms (SNP) were selected from over 57,000 on the Illumina BovineSNP50™ chip for use in genomic evaluations. The discarded SNP were either problematic to score, had a minor allele frequency of less than 2 percent, or were highly correlated with other SNP based on genotypes of 5,503 bulls. Using genotypes from 12,591 animals, comparison of SNP that were homozygous in both parent and progeny found few conflicts when the pedigree relationship was correct, but usually produced over 1,000 discrepancies when it was wrong. The selected genotypes provide a reliable basis for genomic selection.
Technical Abstract: Nearly 57,000 single nucleotide polymorphisms (SNP) on the Illumina BovineSNP50™ chip were investigated to determine their usefulness for genomic prediction. Genotypes were obtained for 12,591 Holstein bulls and cows and SNP selection was done using 5,503 bulls that had genotypes on a larger set of SNP. There were 9,785 SNP that were monomorphic or unscorable and another 3,649 with minor allele frequency <2% which were deleted. Number of SNP for each minor allele frequency class (>= 2%) was fairly uniform (777 to 1,004). For five SNP assigned to chromosome 7, no bulls were heterozygous, which indicated that these SNP are actually on the non-pseudoautosomal portion of the X chromosome. Another 178 SNP that were not assigned to a chromosome that also had many fewer heterozygotes than expected also were assigned to X. Existence of Hardy-Weinberg equilibrium was investigated by comparing observed to expected heterozygosity. Eleven SNP had their expected percent heterozygous differ from actual by more than 15 and were deleted. For 2,638 SNP, there was a high correlation with the genotype of another SNP (genotypes the same for > 99.5% of bulls), and they were deleted. In most cases these were adjacent SNP, or one was unassigned, but 9 were assigned to different chromosomes by the Btau 4.0 sequence assembly. These edits left 40,874 SNP. A parent-progeny conflict was declared when the genotypes were opposite homozygotes. When the pedigree relationship was correct, the mean number of conflicts was 2.3 and when incorrect was 2411. The sire was genotyped for over 93% of the animals. The maternal grandsire’s genotype was similarly checked, but because opposite homozygotes could be valid, up to 16% of cases where both animals were homozygous were accepted. Scanning for identical genotypes detected 3 members of a clonal family, 3 sets of split embryos, 5 sets from ET and 7 identical twins. Genotyping consistency was investigated for 21 bulls genotyped twice with differences being primarily from SNP that were problematic to score. Concordance for readable SNP was extremely high, ranging from 99.96% to 100%. Thousands of SNP that were polymorphic In Holstein were monomorphic in the Jersey or Brown Swiss breeds indicating breed specific sets of SNP are required, or all breeds need to be considered in the selection. The genotypes from the Illumina chip are of high accuracy and provide the basis for genomic evaluations in North American Holsteins.