1	How the genomic evaluation program works
2	Short history Illumina BovineSNP50™ BeadChip developed Accuracy of genomic information assessed by using 2004 evaluations of bulls born before 2000 to predict 2009 evaluations of young bulls Unofficial genomic evaluations of bull calves provided to industry beginning in April 2008 Jersey results released in October 2008 New results released every 2 months Nearly 23,000 animals genotyped through Mar. 2009
3	What is a SNP? Single-nucleotide polymorphism Place on the chromosome where animals differ in the nucleotides (A, C, T, or G) they have Usually not part of the gene that controls a trait – quantitative trait locus (QTL) With enough SNPs, association between SNP alleles and QTL alleles gives useful evaluations SNPs chosen to be distributed evenly and have both alleles well represented in population
4	Source of genomic evaluations DNA extracted from blood, hair, or semen ~40,000 genetic markers (SNPs) evaluated For each SNP, difference in PTA estimated between animals with 1 allele compared to the other allele Genomic evaluation combines SNP effect estimates with existing PA or PTA Genomic data contribute ~11 daughter equivalents to reliability
5	SNP edits and counts
6	How to get animals genotyped Participating AI organizations have 5-year exclusive right to evaluate bulls genomically Each AI organization genotypes first-choice flushes, thereby usually avoiding duplicate genotypes Web-based system collects nominations Avoid duplication Confirm validity of ID and pedigree Associate sample ID with animal ID Breed associations offer cow genotyping service
7	Steps to prepare genotypes Nominate animal for genotyping; confirm not already genotyped Collect hair, blood, or semen from animal Blood not suitable for twins Send to laboratory for extraction Transfer DNA to BeadChip (12 samples/chip) for 3-day genotyping process
8	Steps to prepare genotypes (cont.) Read red/green intensities from chip Transfer intensity files to AIPL for calling genotypes Check genotypes for duplicates, parent-progeny conflicts, and wrong sex
9	DNA laboratories Research Bovine Functional Genomics Laboratory (BFGL), USDA (Beltsville, MD) University of Alberta (Edmonton, AB, Canada) University of Missouri (Columbia, MO) Illumina (San Diego, CA) Commercial (some do extraction only) GeneSeek (Lincoln, NE) Genetics & IVF Institute (Fairfax, VA) Genetic Visions (Middleton, WI) DNA LandMarks (Saint-Jean-sur-Richelieu, QC, Canada) Maxxam Analytics (Mississauga, ON, Canada) ABS (DeForest, WI, through SyGen/PIC, Franklin, KY )
10	What can go wrong Sample doesn’t provide adequate DNA quality or quantity Genotype has many SNPs that can’t be determined (90% call rate required) Genotype conflicts with parent(s) Pedigree error Sample ID error Laboratory error Genotype checked against all others to find true parent
11	Accurate evaluations Accurate genomic evaluations require estimates of SNP effects Evaluations with high reliability provide the most information Recent animals are more useful than ones from earlier generations Reliability of genomic evaluations increases with number of predictor animals
12	Genomic evaluation & reliability Calculate parent average (PA) based only on genotyped animals with best linear unbiased prediction Combine traditional PA (or evaluation) with genomic PA and evaluation using selection index weights Update traditional evaluation with additional information from genomics Reliability from inverse of genomic relationship matrix
13	Data & evaluation flow
14	Genomic vs. traditional PTA Genotype can be thought of as source of information like parents, progeny, and records Official PTA that include a genomic contribution are identified One genotype used to calculate genomic evaluations for all 29 traits Genomic evaluations used the same as traditional PTA Expected to increase rate of genetic improvement because of a large decrease in generation interval
15	Genomic vs. traditional (cont.)
16	Genomic vs. traditional – protein PTA
17	Genomic vs. traditional – net merit
18	Genomic vs. trad. – protein reliability
19	Genomic vs. trad. – net merit reliability
20	Reliability frequencies – young bulls
21	Collaboration with Canada Semex Participated since beginning of genomics research Contributed genotypes to providing a important increase in accuracy for first test Genotypes will be shared between AIPL and Canadian Dairy Network AIPL and University of Guelph collaboration
22	Collaboration with Canada (cont.) Same set of predictor animals used in Canada and U.S. so that evaluations of genotyped animals have same accuracy Canada expects official release of genomic evaluations in August 2009 Common procedures between 2 countries assist in industry acceptance
23	Use of genomic evaluations AI organizations determine which young bulls to buy Considered in selection of mating sires Impact on bull dam selection will increase Used to market semen from 2-year-old bulls
24	January 2009 Genomic evaluations became official Genotyped ancestors contribute their evaluations to descendants Evaluations of all genotyped females are public Evaluations of males enrolled with NAAB or ≥24 months old are public Young-bull genomic evaluations may be shared among AI organizations or disclosed by owner
25	Updates between official evaluations Genomic evaluations calculated approximately every 2 months Evaluations of animals that already have an official evaluation not released Evaluations of new animals distributed to owners Females by breed associations Males by NAAB Usually 1,000–2,000 new genotypes included
26	Distribution of evaluations Nomination establishes a requester who receives the genomic evaluation Requesters 7 participating AI organizations U.S. and Canadian Holstein associations American Jersey Cattle Association Some laboratories Requesting AI organization can agree to share an evaluation with other AI organizations
27	Distribution of evaluations (cont.) Evaluations of all females sent to respective breed associations for distribution to owners NAAB distributes bull evaluations to owners and manages sharing of evaluations among AI organizations Genomic evaluations of animals with official evaluations released as unofficial at updates between official evaluations
28	Impact on producers Young-bull evaluations with accuracy of early 1st-crop evaluations AI organizations marketing genomically evaluated 2-year-olds Bull dams likely to be required to be genotyped Rate of genetic improvement likely to increase by up to 50% Progeny-test programs changing
29	Schedule Calculate SNP effects with each of 3 annual traditional evaluations Calculate genomic evaluations once or more between traditional evaluations Recalculate SNP effects if significant number of predictor animals added May use existing SNP effects if only young animals added
30	Improvements Require bar codes on sample containers to reduce errors and improve lab efficiency Require animals be enrolled with breed association before DNA sample collected Process genotypes frequently; check for and report conflicts as received Reduce processing time by improving efficiency of genotype calling either by laboratories or at AIPL
31	Calling genotypes Scanner reads chip recording intensities of red and green Software converts those to AA, AB, or BB Genotype is missing if assignment is uncertain Accuracy can be improved by adjusting for variation in intensity due to SNP and animal Techniques to automate adjustment are underway Manual intervention can increase accuracy of calling with current software
32	Plans to increase accuracy Genotype more predictor bulls Automatic increase as bulls in waiting receive traditional evaluations Increase number of SNPs used Reach 1,500 Brown Swiss through foreign collaboration? Increase genotyped Jerseys from both domestic animals and possible foreign collaboration
33	International implications All major dairy countries investigating genomic selection Interbull meeting January 2009 discussed how genomic evaluations should be integrated AI organizations need to find balance between competitive benefits from treating genotypes as proprietary versus sharing Importing countries must change rules to allow for genomically evaluated young bulls
34	Longer-term possibilities Determine inheritance of individual chromosome segments (haplotyping) May allow better tracking of QTL Approximate genotypes of missing ancestors to increase predictor population Increase number of SNPs or even use entire DNA sequence
35	Implications Extraordinarily rapid implementation of genomic evaluations Young bull acquisition and marketing now based on genomic evaluations Genomic evaluations may allow more cows from commercial herds to be used as bull dams
36	Financial support National Research Initiative grants, USDA 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) Holstein Association USA (Brattleboro, VT) American Jersey Cattle Association (Reynoldsburg, OH) Agricultural Research Service, USDA