Whole genome selection and the 2000 bull project at USMARC

1
Whole genome selection and the 2000 bull project
at USMARC

• Larry Kuehn
• Research Geneticist

2
DNA Technology

• Genome is highly complex
• Difficult to understand, manipulate
• Historically, oversold to producers
• Need patience, research support, partnerships

3
Definitions

• Genome is the whole set of DNA
• Genetic Markers track regions of the genome
• May be linked/explain phenotypes we observe
• A SNP is a type of genetic marker
• Single Nucleotide Polymorphism
• A Genotype is any pair of markers/marker groups

4
Whole Genome Selection (WGS)

• Use genotypes of thousands of genetic markers,
  like SNP, to predict breeding values (EPDs)
• Like marker sets for traits being offered now but
  with much more density
• Genetic differences in DNA that cause phenotypic
  differences likely close to many markers
• As of now unproven (but preliminary analyses look
  promising)

5
Problem

• Whole-genome selection seems to be big leap
  technology that needs to either be implemented in
  a big way or not at all.
• Early adopters are likely to receive little
  benefit until a critical mass of influential and
  extensively phenotyped animals is tested and the
  data is analyzed jointly through NCE.

6
Whole Genome Selection

• So is now the time?
• We think so!
• Multiple reasons why the time is right

7
Reason 1 Technology
Illumina Infinium Bovine BeadChip

• 50,000 SNP markers across the bovine genome
• On average SNP every gt60,000 base pair
• Discovery SNP includes many breeds

BARC (ARS) USMARC University of
Missouri University of Alberta
8
Reason 2 Data/Cattle

• About 3,000 hd of pedigreed animals with
  extensive phenotypes at USMARC genotyped using
  the 50K chip
• Primarily descendants of Germplasm Evaluation
  (GPE) Cycle VII
• Angus, Hereford, Red Angus
• Charolais, Gelbvieh, Simmental, Limousin

9
Reason 2 Data/Cattle

• Some phenotypes difficult to measure in industry
  setting
• Feed intake
• Carcass data
• Slice shear force
• Structure
• Puberty and productivity
• Disease diagnostic records

10
Reason 3 People

• Collaborative effort between researchers and
  breed representatives to implement WGS for the
  industry

11
2000 Bull Project

• Collaborative effort
• DNA samples from influential or upcoming bulls in
  16 most prominent breeds with a genetic
  evaluation
• Running on 50K chip
• Extensively phenotyped animals from USMARC used
  as training data set
• Establish relationships of markers to phenotypes
  and use results to predict potential of 2000
  industry bulls

12
2000 Bull Project

• Objectives
• Extend genetic predictions from USMARC phenotypes
  to industry bulls
• EPDs developed and delivered to breed
  associations (they decide use)
• Hopefully leads to further bull genotyping
• Validate the effectiveness of WGS using EPDs from
  the 2000 bulls relative to USMARC data on common
  traits (e.g., weaning weight)
• Improve accuracy of EPDs for common traits

13
2000 Bull Project

• Breed associations responsible for selecting
  sires
• High accuracy sires (verify the process)
• Influential sires (greater gain from novel
  traits)
• Number of bulls from each breed dependent on
  breed size
• Maximum of 400 sires from largest breed

14
Potential pitfalls

• We dont know the process of WGS will work
• May need more than 50,000 markers
• Need to develop complex computational methods
• 1992-2006 1.7 million genotypes at USMARC
• 2007-2008 300 million genotypes expected
  from chip results

15
Potential looks good

• Simulated accuracy of genetic prediction

16
SNP associations
Birth Weight
Weaning Weight
Milk
Postweaning gain
Ribeye area
Marbling score
17
Contributors
USMARC -- Tim Smith -- John Keele -- Mark
Allan -- Warren Snelling -- Mark Thallman --
Gary Bennett -- Larry Kuehn -- Tara McDaneld
Bovine SNP Consortium -- ARS, BARC -- U.
Missouri -- U. Alberta -- ARS, USMARC
NBCEC
Breed Associations
ARS NPS -- Steve Kappes