GENOMIC SELECTION IN IRELAND - FROM ZERO TO HERO IN 1 YEAR - DONAGH BERRY1, ANDREW CROMIE2, NOIRIN MCHUGH1 & FRANCIS KEARNEY2
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Genomic selection in Ireland –
from zero to hero in 1 year
Donagh Berry1, Andrew Cromie2,
Noirin McHugh1 & Francis Kearney2
1Teagasc, Moorepark, Ireland
2Irish Cattle Breeding Federation, Ireland
1Dairy and beef industry in Ireland
• Approx. 1.15 m dairy cows and 1.15 m
beef cows
– Dairy - 95% Holstein-Friesian; crossbreeding
on the increase
– Beef – traditional beef cow was 50% dairy
• Several breeding companies (indigenous
and international)
• One phenotypic and genomic database
• One national genetic evaluation center
• Very close working relationship between
farmers, industry, and research
2Before we go any further
• Genomics is USELESS without the basics!!
– Pertinent breeding objective
– Live and accurate phenotypic database
– World class breeding program
• Where are the best cows in the UK for the UK??
3Genetics 101
• Everyone has DNA which is unique
(except for identical twins)
• DNA remains the same throughout life
• DNA clumps together to make genes
• Genes/DNA, interacting with the
environment determine whether an
animal will grow fast or milk well
• Genetic markers (SNPs – “snips”) are
tiny pieces of DNA which can be
measured easily in a laboratory
4Genomic selection
• The theory behind marker assisted
selection was perfect
– Difficult and expensive to find pieces of DNA
causing explaining all the differences in
performance
• Genomic selection is just a large scale
version of marker assisted selection
– If scatter enough measurable genetic markers
across the genome then most of the variation
could be explained
– Genomic tools available (Jerry Taylor yesterday)
– Statistical tools are following
(http://genomicselection.net)
56
7
8
Genomic selection
1. Determine the best DNA signature for
the production system under investigation
– Require the DNA signatures and accurate
estimates of genetic merit for many thousands
of animals
2. Take a hair/blood/tissue sample a young
animal
3. Send off to a laboratory to determine its
DNA signature
4. Compare the DNA signature of the animal
to the best DNA signature for Ireland/UK
9Genomic selection in Ireland
10Timeline in Ireland
Received €2,000 from Irish
Dept. of Agric. (Genetic
conservation Dept.) to store
semen residing at Dept. of
Agric. offices
Mar 07
11Timeline in Ireland
Received funding from the Irish
Dept. of Agric. to genotype a
further 192 Holstein-Friesian
sires for QTL detection for
susceptibility to bovine
tuberculosis
Mar 07
Nov 07
12Timeline in Ireland
•Started to purchase semen from
all AI bulls at the major AI
centers in Ireland (some semen
was donated)
•Start of DNA extraction and
development of DNA bank
Mar 07 Jan 08
Nov 07
13Timeline in Ireland
Started genotyping with AROS
Denmark
Mar 07 Jan 08
Nov 07 Feb 08
14Timeline in Ireland
Additional funding from the Irish
Dept. of Agric. to genotype a
further 768 sires for genomic
selection
Mar 07 Jan 08 July 08
Nov 07 Feb 08
15Timeline in Ireland
•Began research into the
statistical analyses of the data
•National media campaign to
purchase semen straws off
individual farmers (€30/straw)
Mar 07 Jan 08 July 08
Nov 07 Feb 08 Aug 08
16Timeline in Ireland
Undertook first analysis with 803
Holstein-Friesian sires
Mar 07 Jan 08 July 08 Nov 08
Nov 07 Feb 08 Aug 08
17Timeline in Ireland
•Swapped genotypes with LIC in New
Zealand
•Undertook collaborative research in
New Zealand on genomic selection
Mar 07 Jan 08 July 08 Nov 08
Nov 07 Feb 08 Aug 08 Dec 08
18Timeline in Ireland
•Presented accuracy and bias of
genomic selection to Irish dairy
industry
•Presented first unofficial proofs of
genomically selected young bulls
Mar 07 Jan 08 July 08 Nov 08 Jan 09
Nov 07 Feb 08 Aug 08 Dec 08
19Timeline in Ireland
• Genomic proofs became official
• Second country in the world
Mar 07 Jan 08 July 08 Nov 08 Jan 09
Nov 07 Feb 08 Aug 08 Dec 08 Feb 09
20Timeline in Ireland
•Identified and genotyped moderate
to high reliability natural mating
bulls and cows
•Swapped some more genotypes
•Started DNA extraction from beef
bulls
Mar 07 Jan 08 July 08 Nov 08 Jan 09 Sep 09
Nov 07 Feb 08 Aug 08 Dec 08 Feb 09
21Timeline in Ireland
Included swapped genotypes of
international bulls with INTERBULL
proofs but no daughter in Ireland
Mar 07 Jan 08 July 08 Nov 08 Jan 09 Sep 09
Nov 07 Feb 08 Aug 08 Dec 08 Feb 09 Jan 10
22The politics!
• The industry owns the genotypes
• ICBF can use swapped genotypes in
the research and application of
genomic selection
Genomic data
Phenotypic data
ICBF
Genomic breeding values database
Software
23The politics!
• ICBF will provide a genomic
evaluation for any animal from
anywhere in the world if given
access to the genotype which
remains within the ICBF database
– Indigenous AI centers
– International AI centers (irrespective
of where based)
– Farmers (cows or bulls)
24Zero to hero
or
Hero to zero???
25Testing genomic selection
• Split genotyped proven bulls into old
(~75%) and young (~25%)
• You know the genetic merit of all animals
• Estimate the best DNA signature in
the old animals and genetically select
the young animal
• Compare predictions based on
genomics with predictions from old,
accepted system of genetic evaluation
26Genomic prediction - correlations
Trait r Trait r
Milk yield 0.77 Carcass wt 0.35
Fat yield 0.73 Carcass conf 0.68
Protein yield 0.77 Carcass fat 0.74
SCC 0.60 Cull cow 0.74
Fertility 0.73 Locomotion 0.84
Survival 0.58 Overall feet & legs 0.54
Direct calv. Diff 0.37 Overall mammary 0.72
Mat. Calv. Diff 0.56 Overall type 0.72
Gestation 0.62
Peri. Mortality 0.23
27Genomic prediction - bias
Trait Bias Trait Bias
Milk yield 35.84 Carcass wt -6.95
Fat yield -0.36 Carcass conf -4.72
Protein yield 0.40 Carcass fat -0.55
SCC 0.02 Cull cow -0.03
Fertility -1.11 Locomotion -0.04
Survival 0.60 Overall feet & legs -0.16
Direct calv. Diff -0.22 Overall mammary -0.38
Mat. Calv. Diff 0.38 Overall type -0.34
Gestation -0.12
Peri. Mortality 0.40
28Impact on reliability of young bulls
■Parental average; ▲Genomic
0.8
0.7
On average equivalent to 40 daughters
0.6
Some bulls equivalent to 154 daughters
Reliability
0.5
0.4
0.3
0.2
0.1
0.0
l f f t . t
ilk C
C
lity
iv a i f if
i o n
li ty
g h r m Fa io n pe
M S rti rv v. D v. D tat rta e i f o ss ot
l l ty
Fe Su Ca
l al es
M
o s
w
c on
c a om ra
C G f s r c e
ir. at.
a l c a s s C
a
Lo O
v
D M C ar c a
C a
29 ar
CImplementation
of genomic selection in
Ireland
30Implementation
• Active bulls list
– Top 75 reliable (R2>58%) AI sires ranked
on total merit index (EBI) with semen
available
• Should genomically selected animals
be marketed as teams (NZ, NLD) or
as individuals (US)
• Decision was to market as individual
bulls with recommendation to use as
part of a team
31Implementation
• Bulls with GEBVs were eligible for
Active Bull list where reliability for EBI
≥35% and calving proof based on
progeny was ≥50%
– “Layoff bulls”
• Imposed limits on use of semen
–Active Bull List
Bull Details EBI & Proof Details EBI Sub Index
Rk Name of Bull EBI Rel Range Proof Milk FertilityCalving Beef Health
1 O-BEE MANFRED JUSTICE €250 91% +/-€35 DP-IRL €114 €86 €49 -€6 €7
2 RAMOS €216 74% +/-€60 DP-INT €47 €120 €38 -€6 €16
3 HAZAEL LIGHT DETECTOR €198 43% +/-€89 GS €84 €114 €25 -€23 -€2
4 BALLIVOR OLYMPIC GOLD ET €194 50% +/-€83 GS €127 €32 €40 -€5 €1
5 BALLYDEHOB JUSTICE €189 53% +/-€81 GS €90 €76 €28 -€7 €1
6 HILLSDALE LIONEL €188 57% +/-€77 GS €71 €68 €50 -€1 €1
7 MONAMORE ROMERO ET €187 54% +/-€80 GS €90 €72 €38 -€16 €4
8 GIBOR €186 68% +/-€68 DP-INT €74 €81 €25 -€9 €15
9 GARRYMARTIN KEET €184 49% +/-€84 GS €97 €71 €31 -€9 -€6
10 HAZAEL MN SWEETDREAM* €182 35% +/-€95 GS €114 €71 €17 -€12 -€8
DP-IRL – daughter proven with IRL daughters
DP-INT – daughter proven with no IRL daughters
GS- Genomically Selected bulls (Lay-off with calving evaluation ≥50%)
33Active Bull List
2002 2004 2006 2008 2009* 2009**
EBI €72 €103 €113 €124 €127 €150
Average
Reliability
73% 67% 73% 76% 77% 58%
Domestic Bulls 14 12 18 17 24 37
New Top Bull 32 37 11 11 15 48
*Without genomic selection
**With genomic selection
34Uptake
of genomic selection in
Ireland
35Uptake of genomic selection
• 462,000 inseminations from January to November
2009
• Three categories of bulls
– DP-IRL: bulls with daughters producing in Ireland
– DP-INT: bulls with daughters producing ins some country
– GS: genomically selected bulls with a calving survey
Proof No. Bulls Straws/bull % Use Bulls/Herd
DP-IRL 942 177 37 2.7
DP-INT 522 262 29 3
GS 98 1613 34 4
36Why such an uptake?
• Mean difference between DP-IRL and GS
bulls was €69 (more than one standard
deviation better)
• Competitive pricing of GS bulls – range
from €15-€18 euro
• AI companies marketed teams of bulls
(high fertility, high protein, easy calving)
• Discount offered for buying bulls in teams
37Preliminary results (n=35)
Correlation Bias
Parental Genomic Parental Genomic
average average
Milk yield 0.63 0.65 77 50
Fat yield 0.40 0.57 3.0 2.0
Protein yield 0.53 0.65 2.2 1.5
38Genomic selection and
the farmer/breeder
3940
Laboratory
DNA extraction &
genotyping
Improving knowledge of
the best DNA signature
ICBF
Genetic Evaluation
Tag
company
41Genomic selection in
the UK
42What do UK breeders need to do
• Genomic selection will work, the
benefits of which will improve with
time
• The knowledge is out there; UK
(SAC) are heavily involved in
software development
• Collaborators are willing to share
genotypes and ideas
43What do UK breeders need to do
• Industry must agree that genomics
has a role in UK breeding and must
DRIVE the research
– Role of BCBC???
• The UK do not need to be at the
cutting edge but like Ireland be
ready to exploit the work of others
• Leave the politics at the door
44What do UK breeders need to do
• Generate a DNA bank with >10 semen
straws from each dairy an beef bull
• Collate funds (~£1 m) for semen
procurement and genotyping (dairy and
beef)
– Joint investment from industry good bodies
• Implement
• The time is NOW!
45Conclusions
• Genomic selection is a tool to help in
more accurately identifying genetic
elite animals
• Genetic evaluation systems, breeding
objectives and breeding programs
are more important!!
• Genomic selection is very new and we
don’t know will it consistently work…
– But we think it will!!!
46You can also read
