Genetic evaluation programs and future opportunities jueves ingles/James Ro… · Genetic evaluation programs and future opportunities James Rowe (Sheep CRC, Australia) Raul Ponzoni

Genetic evaluation programs

and future opportunities

James Rowe (Sheep CRC, Australia)

Raul Ponzoni (Universidad de la República)

Daniel Brown (Sheep Genetics, Australia)

Julius van der Werf (UNE, Australia)

10th World Merino Conference 2018, Montevideo

Genetic evaluation

estimating genetic merit (breeding values)

NOT – what sort of sheep to breed

NOT – what sort of sheep to produce

Genetic gain determined by:

accuracy of estimating genetic merit

generation interval

0

0.2

0.4

0.6

0.8

1

0 0.5 1 1.5 2 2.5 3 3.5

Accura

cy

Age (years)

Assumedheritability=25%;Accuracyofgenomictest=50%

Performancerecord

Progeny

ParentEBVs

Noperformancerecords

ParentPerformance

Weighing the fleece (George Lambert 1921, Wanganella)

Average fleece weights – Australia 1860-2010

0.0

1.0

2.0

3.0

4.0

5.0

6.0

1860 1910 1960 2010

Ave

rag

e fle

ece

we

igh

t (k

g/h

ea

d )

Trait trends in Australian Merinos (Swan et al. 2017)

Fining the clip FD

YWTIncreasing meat income

CFWFocus on fleece weight

MPP (Mer)

2000 2005 2010 2015

0

40

80

Year of birth

Contr

ibutio

n to in

dex

gain

(%

)

YWT

AWT

EMD

WEC

NLW

CFW

FD

SS

Evolution of Sheep Genetics genetic evaluation

Estimating genetic merit (breeding values)

Pedigree

Performance

Genotype (DNA)

Estimated Breeding Values

(ASBV)

Indexes

Rate of Genetic Gain (index trends)

Maternal

1990 1995 2000 2005 2010 2015

0

1

2

3

4

5

Year of birth

Index tre

nd (

SD

)

MATDOL (BL) MATDOL (CM)

Merino

1990 1995 2000 2005 2010 2015

0

1

2

3

4

5

Year of birth

Index tre

nd (

SD

)

MPP (Mer)

Terminal

1990 1995 2000 2005 2010 2015

0

1

2

3

4

5

Year of birth

Index tre

nd (

SD

)CPLUS (Term)

Swan et al., 2017 AAABG

Maternal Merino Terminal

Ra

te o

f g

ain

–In

de

x tre

nd

(S

D)

2000 2005 2010 2015

0.0

40.0

80.0

120.0

160.0

200.0

1 2 3 4 5 6 7 8 9 10

Cum

ula

tive N

et

Pre

sent V

alu

e

($1000 u

nits)

Years

Faster genetic gain drives profit(Extra net income per 2,000 ewes) (Granleese 2018)

Stud gains(Index points/yr)

6

4

2

1

Genetic evaluation is a key tool

- helps achieve rapid genetic gain

- contributes to well-balanced genetic gain

- but…... expensive

Performance recording

Reference flocks

R&D of the genetic evaluation system

Database management and computing

Costly development of analytical tools

– Single step, MateSel, RamSelect, Flock profiling

Strong case for International collaboration

Competing against other breeds & species –

not against Merino breeders in other countries

Cloud computing makes data sharing easy

Compelling economies of scale in genomics

Standardised DNA testing in multiple countries

Good examples in dairy and beef breeding

G x E concerns increasingly well understood

Shared access to tools (Single Step, MateSel,

RamSelect, Flock Profiling …)

MERINOSELECT evaluation for

Australia and New Zealand (Brown & AGBU)

G x E interactions ?

13

Studied a range of traits – many environments

Accounted for sire by flock & year (SxF) interaction

Conclusions

All traits investigated had high genetic

correlations when Sire x Flock interaction

included

Breeders can select on MERINOSELECT

ASBVs regardless of the country of origin

MERINOSELECT is ‘open’ to concept of

hosting single international evaluation for

Merinos.✔ ? ? ? ?

Tools for improved genetic gain

MateSel available to Sheep Genetics client’s to

help with mate selection.

SingleStep evaluation analysis incorporating:

pedigree, performance & genomics

RamSelect.com.au a web-based app to help

identify rams for specific breeding objectives

Genomic Flock Profiling average flock breeding

values from DNA testing 20 latest drop progeny.

A benchmark to guide ram purchases.

Opportunities

Analysing genetic gain (From Swan et al 2017)

How does actual gain for Merinos

compare to potential gain?

How do individual ram breeders

compare?

Actual gain as % of potential gain(Swan et al. 2017)

Maternal

2000 2005 2010 2015

25

50

75

100

Year of birth

% p

ote

ntial gain

MATDOL (BL) MATDOL (CM)

Merino

2000 2005 2010 2015

25

50

75

100

Year of birth

% p

ote

ntial gain

MPP (Mer)

Terminal

2000 2005 2010 2015

25

50

75

100

Year of birth

% p

ote

ntial gain

CPLUS (Term)

Maternal Merino Terminal

100

75

50

25

2000 2005 2010 2015 2000 2005 2010 2015 2000 2005 2010 2015

Comparing gains for individual breeders)

-20

0

20

40

60

80

100

120

140

MPP (Mer) CPLUS (Term)

Top

20%

Bottom

20%

Top

20%

Bottom

20%

Merino Terminals

(From Stephen et al 2018)

Full pedigree data is one problem

0

20

40

60

80

100

Merino Terminal

(From Stephen et al 2018)

Top

20%

Bottom

20%

Top

20% Bottom

20%

Merino Terminals%

of a

nim

als

with

fu

ll p

ed

igre

e

Genomic Selection: the big picture

21

(Photo: Julius van der Werf 2018)

Information Nucleus – innovation platform

Information

Nucleus

DNA

MeatWoolSheep

Phenotype data

1. Understanding complex phenotypes

2. Quantifying G x E

3. Genomic prediction of breeding values

4. Bio-bank (DNA and database)

Genomics

207Blending GBLUP EBVs with

ASBVs (2012)

Single Step Carcase Analysis

(2016)

Full Single step in Main Analyses

(2017)

Impact on industry through genomics (genetic gain - index points/year)

2000-2010 2011-2017 Difference

Merinos (MP+) 1.57 2.19 +39%

Terminals (C+) 3.85 4.29 +11%

Terminals (LEQ) 1.36 2.00 +47%

Some Confounding factors

e.g. Index development &

Reference population(Brown et al 2018)

Prediction accuracy: Meat Traits in Merino

0.0

0.1

0.2

0.3

0.4

0.5

0.6

ccfat cemd imf pemd sf5 pwt

Pre

dic

tio

n A

ccu

racy

50K 50K+Top Seq (2)

Value of genomicsearly information and difficult to measure traits

0

0.2

0.4

0.6

0.8

1

0 0.5 1 1.5 2 2.5 3 3.5

Accura

cy

Age (years)

Assumedheritability=25%;Accuracyofgenomictest=50%

Performancerecord

Progeny

ParentEBVs

Noperformancerecords

ParentPerformance

Genomic

Conventional

DNA tests getting cheaper and predictions more accurate

Signs of rapid change

y=34.45x-69117

-

50

100

150

200

250

300

350

400

2004 2006 2008 2010 2012 2014 2016 2018

Numberofstudsregisteredin

MER

INOSELCT

Increase in membership of MERINOSELECT

Rapid increase in poll ram semen sales(Note: DNA test developed 2009)

0

4000

8000

12000

16000

2005 2007 2009 2011 2013 2015

Dose

s s

old

(N

SW

AA

SM

B)

Increase in poll ram sales (‘Top 20’ NSW studs)

(Note: DNA test developed 2009)R

am

s s

old

(N

SW

AA

SM

B top

20)

0

2000

4000

6000

8000

10000

2005 2007 2009 2011 2013 2015 2017

Poll

Horn

-

20,000

40,000

60,000

80,000

2010 2012 2014 2016 2018

(a) DNA parentage test numbers

(per year)

-

5,000

10,000

15,000

20,000

2010 2012 2014 2016 2018

(b) Genomic test numbers (per year)

Increasing use of DNA (Genomic) testing

Concluding comments

• Genetic evaluation programs are crucial for rapid and well-

balanced genetic gain

• Many Merino breeders can achieve much faster genetic gain for

a range of traits required by their clients

• Genomics offers huge potential for Merinos

• New tools and services (Single Step, MateSel, RamSelect, Flock

Profiling) assist in making best use of genetically superior sheep

• International collaboration – a strong case a single evaluation

program based on MERINOSELECT

? ? ? ?