1. Bull Selection Strategies Using Genomic Estimated Breeding Values L. R. Schaeffer CGIL, University of Guelph ICAR-Interbull Meeting Niagara Falls, NY June 18, 2008

2. Understanding Cancer and Related Topics Understanding SNPs and Cancer These PowerPoint slides are not locked files. You can mix and match slides from different tutorials as you prepare your own lectures. In the Notes section, you will find explanations of the graphics. The art in this tutorial is copyrighted and may not be reused for commercial gain. Please do not remove the NCI logo or the copyright mark from any slide. These tutorials may be copied only if they are distributed free of charge for educational purposes. Developed by: Susan Greenhut, M.S. Donna Kerrigan, M.S. Jeanne Kelly Brian Hollen Explains tiny variations in the human genome called Single Nucleotide Polymorphisms (SNPs) that can influence a person’s health. Shows how SNPs occur in both coding and noncoding regions and can cause silent, harmless, harmful, or latent effects. Shows how SNPs can be markers for cancer. Suggests that SNPs may also be involved in the different levels of individual cancer risk observed. Suggests that, in the future, SNPs databases may be used to improve cancer diagnosis and treatment planning.

3. What Is the Human Genome? Human Cell Nucleus Chromosomes

4. DNA and Chromosome Structure DNA molecule (chromosome)‏ Chemical bases A T G C

5. The Genome Contains Genes Gene 2 Coding region Protein 2 Protein 1 Noncoding region Gene 1 Coding region

6. Variation in the Human Genome Person 1Person 2 = Variations in DNA

7. What Is Variation in the Genome? Common Sequence Variations Polymorphism Deletions Translocations Insertions Chromosome

8. Variations Causing No Changes = Variations in DNA that cause no changes

9. Variations Causing Harmless Changes = Variations in DNA that cause harmless changes

10. Variations Causing Harmful Changes = Variation in DNA that causes harmful change No Disease Hemophilia

11. SNPs Are the Most Common Type of Variation At least 1 percent of the population Most of the population Common sequence G to C SNP site Variant sequence

12. Step 1: Deduce QTL genotypes * Treat markers as QTLs LD, r2 > 0.35, markers are 50 kb apart 60,000 markers evenly spaced Across breeds – same linkage phase, Holstein – Angus when markers 10 kb apart * Marker genotypes versus haplotypes Can increase proportion of QTL variance Multiple marker genotypes inbetween

13. Step 2. Estimate QTL Effects p(u) approx, exponential high probability of p(0) - QTLs with no effects.

14. Terminology REFERENCE POPULATION 1.DISCOVERY DATA SET - to estimate QTL effects 1.VALIDATION DATA SET - to assess accuracy of GEBV Function or procedure to determine GEBV derived. Cross Validation.

15. SELECTION CANDIDATES PhenotypesPedigreesMarkers Not all will have genotypes (95%) 10K, 50K, others Platforms will keep changing

16. SELECTION CANDIDATES PhenotypesPedigreesMarkers EBV Traditional GEBV I = b1(EBV) + b2(GEBV)

17. SELECTION CANDIDATES PhenotypesPedigreesMarkers Infer genotypes for all animals Efficient algorithm Form Relationship Matrix from Marker genotypes - M BLUP using M (not A) Improved EBV

18. SELECTION CANDIDATES PhenotypesPedigreesMarkers EBV Traditional GEBV Multiple Trait analysis of EBV and GEBV together, covariances? Others???

19. Design and Application of the BovineSNP50. Curt Van Tassell USDA Agricultural Research Service curtvt@ars.usda.gov 301-504-6501

20. History Goals 1.Implement whole genome selection 2.QTL detection Challenge –Needed at least 30,000 good markers Problem –Product did not exist

21. Cattle SNP Collaboration iBMAC Develop 60,000 Bead Illumina iSelect® assay –USDA-ARS Beltsville Agricultural Research Center: Bovine Functional Genomics Laboratory and Animal Improvement Programs Laboratory –University of Missouri –University of Alberta –USDA-ARS US Meat Animal Research Center Starting 60,800 beads – expected 53,000 SNPs to result Plan to genotype ~30,000 animals for multiple projects

22. SNP Available for Assay Design With MAF (18%)‏  Next Generation Sequencing 62,042  Bovine HapMap Consortium33,836  DPI, US-MARC,UA 10,574 InSilico SNPs (72%)‏  Assembly SNP (Filtered)278,429  Baylor Interbreed123,049  BAC and BAC-end Derived89,832  INRA764 Total: 598,64885% Infinium II (1 Bead)‏ 15% Infinium I (2 Beads)‏

23. Gap Distribution 80% gaps < 60kb

25. 1 2 2 2 2 1 1 1 1 2 1 2 2 1 2 2 +.3kg-.1kg 50,000+ interval estimates Haplotypes or Genotypes Hotspots, blocks of markers GEBV = (Sum of estimates -.1kg 0kg Genotypes )‏

26. GEBV 1.Estimate 50,000+ genotypes or haplotypes and sum together estimates given genotypes 2.Use genotypes to get more precise A matrix 3.Use polygenic model and add in significant explanatory markers (100-200 markers)‏ 4.Apply data-mining, machine learning to select 3,000 to 5,000 SNP markers

27. Advantages of GEBV?? Better accuracy than Parent Average EBV Can be obtained at birth of animal

28. Disadvantages?? SNP interval estimates deteriorate over time Effects need to be re-estimated Estimates differ between populations SNP panels will change over time

29. Implications on Bull Selection Assume that GEBV works - Accurate, stable - Available at birth

30. Progeny Testing Scheme (PT)‏ 1 Million cows 400 young bulls per year (100 daus)‏ 50 proven bulls per year Top 20 used as sires of sons Dams from top 5% of population (0.50)‏ Young bulls $10,000 purchase, $6,000/yr Single trait, heritability=0.30

31. PT Time Frame Year 0 – young bull is born Year 1 – test matings are made for young bulls Year 2 – daughters are born Year 3 – daughters are bred Year 4 – daughters calve, first lactations begin Year 5 – young bulls receive first proofs, culled or returned to service, second crop of daughters started 6 years from birth to proof.

32. Years GSD Change 0 1 2155.26, $18.7M Genetic Change

33. A: Use GEBV to select young bulls Young bulls selected using GEBV Still go through usual PT scheme Corr(GEBV,TBV) = 0.4 to 0.8 500 to 4000 young bulls genotyped at birth Purchase price is still $10,000 per bull $300 to genotype an animal

34. Years GSD Change 0 1 2155 PT A: Corr(GEBV,TBV)=0.4 4000, $20.6 2000, $20 1000, $19.6 500, $19.4

35. Scheme A: Corr(GEBV,TBV) Increasing from 0.4 to 0.8, no change in GSD per year – about 0.34. Number of Young Bulls Genotyped More is better, balance against costs Number Chosen for PT (400 – 200 – 100)‏ No drop in GSD, costs decrease to $6.1 M for 100 bulls

36. B: Use Young Bulls as Sire of Sons Genotype 500-4000 young bulls at birth, GEBV Select 50, best 20 as Sires of Sons for next generation Cost of purchasing a bull = $20,000 Special contracts may be needed

37. Scheme B: Time Frame Year 0 – young bull is born AND GENOTYPED, GEBV Year 1 – Use as though a proven sire, sire of sons Year 2 – Year 3 – Year 4 – Year 5 – young bulls receive usual first proof 0 years from birth to proof, no need to wait 6 years

38. Years GSD Change 0 1 2155 PT B: Corr(GEBV,TBV) = 0.4 4000, $6.7, 0.53 2000, $5.1, 0.50 500, $3.1, 0.44

39. Scheme B: Corr(GEBV,TBV) Increasing from 0.4 to 0.8, increase GSD 0.50 to 0.55 for 2000 genotyped young bulls Number of Young Bulls Genotyped More is better, balance against costs - 2000 Turnover rate – 36 out of 50 bulls per year Costs are less than $10 M per year

40. C: Between Schemes A and B Genotype 2000 young bulls at birth, GEBV Select best 200 for usual PT Best 5 or 10 used as Sires of Sons at 1 yr for 1 year Cost of purchasing a bull = $10,000 Corr(GEBV, TBV) = 0.6

41. Years GSD Change 0 1 2155 PT C: Corr(GEBV,TBV) = 0.6, 2000 GEBV Scheme B 2000, $5.1, 0.50 Scheme C, 10, $10.1, 0.46

42. Results Genotype 2000 bulls per year, GEBV Reduce number for PT to 100-200 Use top 10 as Sires of Sons Gradually phase out PT and go to Scheme B Share SNP results with world

43. Genotyping Females Genotype at birth, GEBV Use top 5000 as dams of bulls Use top cows as herd replacements No point in genotyping current dams of bulls – already known entities. GEBV better than EBV – no pref trt.

44. Conclusions Use GEBV to select, and use young animals as though they were proven Costs may increase for young animals Inbreeding should be studied Strategies for producers to be involved International implications (share or not)‏

45. More SNP panels, ever changing SNPs to QTLs (patents?)‏ G x E Breeds