Selection of single nucleotide polymorphisms and genotype quality for genomic prediction of genetic merit in dairy cattle

1
Selection of single nucleotide polymorphisms and
genotype quality for genomic prediction of
genetic merit in dairy cattle
Abstr. 524
2
Bovine SNP50 BeadChip
Bovine HapMap Consortium
3
Laboratories animals genotyped

• Bovine Functional Genomics Laboratory, 8,236
• ARS, USDA (Beltsville, MD)
• GeneSeek (Lincoln, NE) 2,288
• University of Alberta (Edmonton, AB) 1,325
• University of Missouri (Columbia, MO) 914
• Genetics IVF Institute (Fairfax, VA) 703
• Illumina (San Diego, CA) 116

4
Genotypes by breed sex
Breed Female Male All
Brown Swiss 3 348 351
Holstein 1,567 9,046 10,613
Jersey 0 589 589
All 1,570 9,983 11,553
5
SNP filtering counts

• SNP available 58,336
• (including those not supported by Illumina)
• Insufficient average number of beads 1,389
• Unscorable 4,360
• Monomorphic in Holsteins 5,734
• Minor allele frequency (MAF) of lt5 6,145
• Percentage heterozygous different from 282
• expected or other scoring issues
• Highly correlated 2,010
• Used for genomic prediction 38,416

6
Bull distribution for unscorable SNP
4,389 Holstein bulls genotyped
7
SNP distribution for bulls with unscorable SNP
8
Identical twins clones

• 
  Missing SNP Changes
• GLC Triton-ETN 19 10
• GLC Triad-ETN 348 364
• GLC Trey-ETN 14 5
• Comestar Laureat 13 2
• Comestar Loyalty 16 5
• Bermath Monroe 160 156
• Bermath Morgan 83 76
• Granduc Blizzard 13 10
• Granduc Spartan 36 33
• Granduc Performance 8 4
• Granduc Performer 988 984
• U-of-Minn W Fatal Twyla 2,272 2,223
• U-of-Minn W Fatal Tabina 116 65

9
Genotyping for same bulls

• 2 laboratories 46 bulls 38,416 SNP
• SNP unscorable only once
• Mean of 792 SNP
• Range of 20 to 2,244 SNP
• SNP conflict (lt0.003)
• Mean of 0.9 SNP
• Range of 0 to 7 SNP

10
Highly correlated SNP

• 4,286 bulls 40,426 SNP
• Compare each SNP with all others with MAF within
  2
• Count all 9 combinations of 0, 1, and 2 for 2 SNP
• Assumed highly correlated if only combinations
  with gt21 SNP are 00, 11, and 22 or 20, 11, and 02

11
Highly correlated SNP (cont.)

• 2,444 pairs detected and categorized into 1,696
  groups
• SNP in group Groups
• Group sizes 2 1,450
• for highly 3 198
• correlated 4 30
• SNP 5 10
• 6 6
• 7 1
• 10 1

12
Highly correlated SNP (cont.)

• Distance between autosomal SNP
• 65,531 mean base pairs between consecutive SNP
• 19,826 base pairs between 1,199 consecutive
  highly correlated SNP
• 423,033 base pairs between 469 SNP in group where
  ?1 SNP in group is not consecutive

13
Minor allele frequency distribution
38,416 SNP
14
Heterozygosity differences

• SNP different from expected heterozygosity by gt7

Chromo-some SNP MAF () Actual () Expected ()
15 Hapmap39475-BTA-20743 20 39 31
18 ARS-BFGL-NGS-9311        41 72 48
26 ARS-BFGL-NGS-97477 47 95 50
29 ARS-BFGL-NGS-32567 30 59 42
X ARS-BFGL-NGS-74090 42 59 49
BTB-01316369 22 44 34
SNP assigned based on bull homozygosity SNP assigned based on bull homozygosity SNP assigned based on bull homozygosity SNP assigned based on bull homozygosity SNP assigned based on bull homozygosity
15
Sex determination

• 605 SNP on X chromosome and not on Y
• (not in pseudo-autosomal region)
• Bulls should not
• be heterozygous
• Cows (n 255)
• Mean of 229.6 heterozygous SNP
• Range of 133 to 277 heterozygous SNP

16
Parentage verification discovery

• Compare parent and progeny SNP if both homozygous
• If gt200 conflicts found, check every other
  genotype for one with lt30 conflicts
• May be parent, progeny, duplicate, clone, or
  identical twin

17
Estimating missing SNP - imputing

• Base population gene frequencies estimated with
  best linear unbiased prediction
• Estimates provided for all genotyped animals and
  ancestors
• If SNP genotype missing, estimate used if within
  0.2 of 0, 1, or 2
• 6,268 genotyped animals
• 0.9 missing
• 45 filled

18
Summary

• High-density Illumina BovineSNP50 BeadChip
  available
• gt38,000 SNP available for Holsteins
• Genotypes consistent across laboratories
• Few parent-progeny genotyping errors found
• If parentage error found, correct parent often
  can be determined

19
Summary (cont.)

• SNP on X-chromosome support sex validation
• Call rate is high and may be improved
• Detection of highly correlated SNP revealed
  probable errors in map locations

20
Financial support

• National Research Initiative grants, USDA
• Natl. Assoc. of Animal Breeders (NAAB, Columbia,
  MO)
• ABS Global (DeForest, WI)
• Accelerated Genetics (Baraboo, WI)
• Alta (Balzac, AB)
• Genex (Shawano, WI)
• New Generation Genetics (Fort Atkinson, WI)
• Select Sires (Plain City, OH)
• Semex Alliance (Guelph, ON)
• Taurus-Service (Mehoopany, PA)
• Agricultural Research Service, USDA