Multi-breed beef genomics Research update April 2015 Donagh Berry Teagasc, Moorepark Herdbooks, April 2015
Genomic selection Uses DNA information to supplement pedigree information to identify the genetically elite animals DNA is the same throughout life and in each cell of the body Even if you know the genetic merit of parents with 100%, max reliability of progeny is 50% Which half of the genes did you inherit from each parent?
International perspective Implemented in dairy in Ireland in 2009 Up to 54% more accurate 60% of semen sold last year Within-breed genomic predictions in US and Canada Run by Dorian Garrick Black Angus, Red Angus, Limousin Within-breed genomic predictions in France CH, LM, BA official in Autumn
Approach 2 1 4 6 7 8 0.50 0.25 6 7 8 6 1.00 0.50 0.25 7 0.50 1.00 0.25 8 0.25 0.25 1.00 0.25
+20 kg +10 kg 2 1 4 6 7 8 0.50 +15 kg +15 kg +20 kg +10 kg 2 1 4 6 7 8 0.54 +14 kg +16 kg Approach 6 7 8 6 1.00 0.50 0.25 7 0.50 1.00 0.25 8 0.25 0.25 1.00 Genomic Better relationships 6 7 8 6 1.02 0.54 0.22 7 0.54 1.00 0.28 8 0.22 0.28 1.00
Imputation Use 40,000 DNA markers Costs 79 Lower cost of ~14,000 DNA markers Predict (Impute) the remaining 26,000 Genotyping v mobile phones More power but the same price ~14,000 DNA markers 40,000 DNA markers for the same price Beagle Fimpute software Faster and more accurate - cost
Imputation - test Breed Breed N Mean Mean Min Min Max Max AA 2897 0.98 0.59 1.00 AU 61 0.76 0.75 0.79 BA 133 0.78 0.74 0.90 BB 457 0.98 0.86 1.00 BB CH 9335 457 0.98 0.98 0.86 0.59 1.00 1.00 CH FR 9335 1188 0.98 0.96 0.59 0.54 1.00 0.99 FR HE 1188 2285 0.96 0.98 0.54 0.56 0.99 1.00 HE HO 2285 6128 0.98 0.97 0.56 1.00 IM 46 0.87 0.80 0.91 HO 6128 0.97 0.56 1.00 JE 32 0.98 0.95 0.99 IM LM 8903 46 0.87 0.98 0.80 0.72 0.91 1.00 JE PI 32 23 0.98 0.75 0.95 0.72 0.99 0.81 LM PT 8903 172 0.98 0.75 0.72 0.71 1.00 0.97 SA 135 0.79 0.75 0.99 PI 23 0.75 0.72 0.81 SH 98 0.86 0.82 0.95 PT SI 172 1507 0.75 0.98 0.71 0.81 0.97 1.00 SA 135 0.79 0.75 0.99 SH 98 0.86 0.82 0.95 SI 1507 0.98 0.81 1.00
Imputation - test Breed Breed N Mean Mean Min Min Max Max AA 2897 0.98 0.59 1.00 AU 61 0.76 0.75 0.79 BA 133 0.78 0.74 0.90 BB 457 0.98 0.86 1.00 BB CH 9335 457 0.98 0.98 0.86 0.59 1.00 1.00 CH FR 9335 1188 0.98 0.96 0.59 0.54 1.00 0.99 FR HE 1188 2285 0.96 0.98 0.54 0.56 0.99 1.00 HE HO 2285 6128 0.98 0.97 0.56 1.00 IM 46 0.87 0.80 0.91 HO 6128 0.97 0.56 1.00 JE 32 0.98 0.95 0.99 IM LM 8903 46 0.87 0.98 0.80 0.72 0.91 1.00 JE PI 32 23 0.98 0.75 0.95 0.72 0.99 0.81 LM PT 8903 172 0.98 0.75 0.72 0.71 1.00 0.97 SA 135 0.79 0.75 0.99 PI 23 0.75 0.72 0.81 SH 98 0.86 0.82 0.95 PT SI 172 1507 0.75 0.98 0.71 0.81 0.97 1.00 SA 135 0.79 0.75 0.99 SH 98 0.86 0.82 0.95 SI 1507 0.98 0.81 1.00 Genotyping the minority breeds in the coming weeks
Genomic predictions
<1970 1971-1975 1976-1980 1981-1985 1986-1990 1991-1995 1996-2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 Number of animals genotyped Number of animals in the analysis Genotyped animals 20000 18000 16000 14000 12000 10000 8000 6000 4000 2000 0 1400 1200 1000 800 600 400 200 0 Year of birth
<1970 1971-1975 1976-1980 1981-1985 1986-1990 1991-1995 1996-2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 Number of animals genotyped Number of animals in the analysis Genotyped animals 20000 18000 16000 14000 12000 10000 8000 6000 4000 2000 Reference Validation 1400 1200 1000 800 600 400 200 0 0 Year of birth
Units (Scale 1-9) Mean Bias 0.06 0.04 0.02 0.00 Traditional Genomic -0.02-0.04-0.06-0.08-0.10-0.12 Remember the true proof also still has traditional information -0.14
Correlation Correlation 0.80 0.70 0.60 0.50 0.40 0.30 0.20 0.10 0.00 The true proof is still only 48-52% reliability Correlation of >0.95 with proven (i.e., >70% reliability) animals
0.6 Reliability 0.5 0.4 0.3 0.2 0.1 Traditional Genomic Genotypes without phenotype or pedigree without reliability 0.0
Reliability 2013 calves (no records) 0.50 0.45 Traditional Genomic 0.40 0.35 0.30 0.25 0.20 0.15 0.10 0.05 0.00
Reliability No information coming from genotype
Reliability 45% 90% 30% reliability Proof likely to move less with time
95%CI = 1.96 1 R σ Reliability Proportion >95% CI Trait DEVHQ 1.03 HW 2.40 LENB 95% reside inside limits 2.05 LENP 7.68 LOIND 2.70 WBEHW 0.19 THIGH 3.12 WATW 0.49 WAHPS 3.99
Genomic Bull movements with records (HqDev) Traditional Down Stay Up Sum Down 0.22 0.54 0.24 1.00 Stay 0.19 0.48 0.34 1.00 Up 0.14 0.49 0.37 1.00 Of the animals that genomic predictions said would improve relative to parental average, 14% got worse, 49% stayed the same and 37% went up Almost 3 times more likely to increase than decrease
Genomic Bull movements with records (Loin Dev.) Traditional Down Stay Up Sum Down 0.26 0.46 0.29 1.00 Stay 0.17 0.48 0.35 1.00 Up 0.15 0.46 0.39 1.00 Almost 3 times more likely to increase than decrease
Genomic Bull movements with records (length of Back) Traditional Down Stay Up Sum Down 0.29 0.42 0.29 1.00 Stay 0.25 0.49 0.26 1.00 Up 0.19 0.46 0.34 1.00 Almost 2 times more likely to increase than decrease
Conclusions Little or no bias in genetic evaluations No over-prediction of genetic merit of young animals Good lift in reliability 29% 45% 0.5% increase for proven animals Good (early) indicator of animals likely to increase in genetic merit with time
Acknowledgements Research Stimulus Fund MultiGS Breed societies, AI stations and farmers Ismo Stranden and Esa Mäntysaari Beef Genomics Scheme Technical Advisory Group