where next for terminal sire recording?

Post on 11-May-2022

0 Views

Category:

Documents

0 Downloads

Preview:

Click to see full reader

TRANSCRIPT

Where next for

Terminal Sire Recording?

Samuel Boon, Signet Manager

and EBLEX Breeding Specialist

Performance recording has

been a great success......

High rates of genetic gain

for growth and carcase traits

Success

• Breeders are using EBVs to add value

to ram sales

• Increase in demand for recorded rams

• Commercial understanding is increasing

(but still a long way to go)

Number of flocks recording is limited

and hence industry impact could be

higher

Challenges

Breeding Values are getting bigger......

Is it time to rebase?

Challenges

Breeders are looking for more regular

BLUP information – 3 runs a year is

limiting

More crossbred animals and crossbred

flocks are looking for useful EBVs

Current BLUP solutions are limited

Commercial demand.......

At every sheep meeting I have spoken at for 15+

years someone has said “can we compare breeds”

Would an indirect “comparison” – if promoted sensibly

- assist breeders (and breeds) of high genetic merit

sheep?

Should Signet be helping the industry to find the

sheep with the best genes – regardless of breed?.....

We have been here before.....

Current approach In

de

x /

EB

V

1990 2013

Current approach In

de

x /

EB

V

1990 2013

Charollais

Suffolk

Texel

Meatlinc

Across breed

approach In

de

x /

EB

V

1990 2013

Charollais

Suffolk

Texel

Meatlinc

Combined breed evaluations

What are other people doing?

£ £

UK Dairy

Evaluations

“Everyone in one

pot”, but breeds

reported separately

Lambplan

Combined runs

for Merino + Rest

SIL

Central Progeny Test

ICBF

Beef in Ireland

Sheep Ireland

(future?)

Starting with a simple approach

– doesn’t preclude more advanced solutions in the future

Where will we get crossbred

data?

• Existing members with crosses?

• New members with crosses?

• Other sources? Test flocks?

Advantages

(potentially)

• More regular EBV updates

– More efficient use of money to produce

EBVs

• More accurate EBVs for crossbreds

– Greater accuracy values for all

• Clearer message for commercial

producers

• Even greater interest in top end

genetics

Independence

• Within breed rankings remain

• EBVs across breeds

– but breed specific indexes are possible

• Breed specific genetic group solutions

• For those not interested in other

breeds

– not really an issue

– business as usual

Crossbreds - are they

inevitable?

The introduction of Composites has not only

lifted sheep performance by their own merit,

it has also put a sharper focus on breeding

and performance of purebred sheep

Mark Young, Heartland Sheep October 2013

Threat? Opportunity?

Summary

This is a major opportunity to improve

the way we identify and promote

terminal sires with superior genes

Thank you

Combined breed analysis ~ technical issues

Kirsty Moore

19 19

Introduction

• Improved efficiency

– Increased number of evaluations

– Increased accuracies

• Ability to evaluate cross bred animals

• Commercial producer

• Be on par with international methods and systems

• No impact if not interested in benefits

20 20

Issues along the way

• Heterosis

• Genetic parameters

• Breed effects

• Genetic groups

• Rebasing

• Different indices

• Running time

• Publishing results

21 21

The road is partially paved ...

• Combined breed analysis is not a new concept

• UK dairy evaluations have been combined breed

since ~2010

• 2013 EasyRams SPARK award

• Have started a test terminal combined breeds

evaluation

22 22

Some of the easier questions

• Heterosis

• Genetic parameters

• Breed effects

• Genetic groups

• Rebasing

• Different indices

• Running time

• Publishing results

Benchmark EBVs across runs over time. 1990 born animals have an average EBV of 0. Modernise the ‘base animals’

Yes breeds can have different indices depending on their different breeding goals

Yes it will take longer to run but we have the computing power and software to handle this. The dairy evaluation ~ 20 million milk records

The changes proposed concern the calculation of EBVS we can still publish onto BASCO, web search, extract specific groups etc

23 23

Heterosis / Hybrid Vigour

• When cross bred progeny perform better than expected given the 2 parental breeds – Reproduction

– Survival

– fitness

• In this example we expected 55kg but observed 65kg. The extra 10kg is due to hybrid vigour

50kg 60kg 55kg

65kg

24 24

Heterosis / Hybrid Vigour

• But heterosis is not genetic!

• If we ignore heterosis

– estimate that an animal (and its relatives) is better/worse genetically than it really is

• Relatively easy to account for

– Different crosses express different amounts of heterosis

– knowledge of the breed makeup

– Proportion of Each Breed (PEB)

• Based on information from BASCO/breed 16ths

• many breeds to be present

• records breed make up as a %

• Good agreement with existing breed 16ths

• Better able to record breed of complex breed crosses

– Breed or breed type?

25 25

Genetic parameters

• Can only use 1 set

• Review of parameters

– show very similar heritabilities/correlations

– In some cases differences in variances

• Not sure if this a true reflection of the data today

– Need to extract some data and look at the phenotypic

variances to see if there are breed differences

• Not a problem if variances are different as we can model this

• Dairy pre scale data so variances are comparable

26 26

Breed effects / genetic groups

• Genetic group definitions will need to revised to

include breed

– Provide the opportunity to improve the definitions

• We will need to account for breed effects

27 27

EasyRams SPARK award

• Opportunity in a small dataset to start to work with

some of the issues

• Robyn Hulme

– Suffolk, Texel and SufTex

– NZ importations

• 2 parts to the SPARK

– Across country

– Combined breed

28 28

The data

• Data from 2007-2012

– Robyn’s flocks or flocks with a number of progeny from a

All Black ram produced by Robyn

– ~4,500 with data

– ~18,000 5 generation pedigree

• PEB procedures / breed makeup

• Data extraction

– Some modifications required

• Dam breed

• Genetic groups

29 29

The EBVS

• EBVs produced for All Black animals

– 8wk weight

– scan weight

– muscle depth

– fat depth

• Correlation between the combined breed EBV and within breed EBV

– 0.69 – 0.94

– Indicates that heterosis does change the ranking but not radically

– Some of the re-ranking is due to changes in models and data sets

30 30

How did EBVs compare with phenotype

SWT MD FD

R(phen & within breed)

combined breed)

within breed)

combined breed)

within breed)

combined breed)

Suffolk 0.41 0.44 0.54 0.55 0.62 0.62

SufTex 0.48 0.63 0.60 0.63 0.51 0.59

Mostly Tex 0.22 0.39 0.81 0.76 0.63 0.70

n=~1300 animals

31 31

EBVs for 17 sires

-2

0

2

4

6

8

10

12

14

-4 -2 0 2 4 6 8

Wit

hin

Bre

ed

EB

V

Combined Breed EBV

SWT

MD

FD

32 32

Summary

• We have started an combined terminal breed

analysis test run

– Still working out which methods and approaches are

appropriate for our data

• A massive step forward for the UK sheep industry

– Vast improvements in efficiency

– Better service the users

– Improvement in EBVs and accuracy of the EBVs

– Utilising cross bred data

Multi-Breed Genetic Evaluations

Lessons from UK Dairy evaluations

Marco Winters

DairyCo Breeding+

• Responsible for Genetic Evaluation in UK

– Independent and Paid for by dairy farmers

• All breeds and crosses :

– Production traits

– SCC

– Lifespan

– Fertility Index

– Type (excl. B&W)

– Calving Ease

Who do we work with? Breed Societies Milk Recording Service partner

Critical success factors;

• Recording (ICAR accredited)

• Collaboration – (inter)nationally

The Breeders ‘toolbox’

• Dairy breeding has never been so easy !

– Many bulls on offer from many breeds

– Many genetic indexes available to use

• However, they only add value if they are used !

– Regardless of heritability

Yield 1977 – 2006 (Year of Birth) (Milk Genetics vs. 1st Lact. yields)

3000

3500

4000

4500

5000

5500

6000

6500

7000

7500

8000

-3500 -3000 -2500 -2000 -1500 -1000 -500 0 500

HOL

SHO

AYR

JER

GUE

FRI

Impact of Genetics – lower h2

0

50

100

150

200

250

300

-35 -25 -15 -5 5 15 25 35

Dtr

Lacta

tio

n A

vg

. S

CC

Sire PTA

Lact.1

Lact.2

Lact.3

Lact.4

• Daughter average – Lactation SCC

Standardised Genetic Gains (based on insemination data)

Future Challenges - Competitiveness

• What are the future genetic needs ?

– Consider future economic conditions

– Consider different ‘non-economic’ demands

• E.g. environment, welfare, consumer

– Consider ever-widening range of production systems

• What are implications for Genetic evaluations ?

– Are we making best use of available data?

Genetic Evaluations

• Performance = Genetics + Environment

• Genetic evaluations based on:

– Pedigree information

– (Genomic information)

– Performance recording (e.g. Milk, SCC)

• Correcting for environmental effects

– Progeny performance • Proper adjustment for genetic merit of mate

• Genetic gain improves with higher accuracy

– (but there is a trade-off with Generation Interval)

Time & Accuracy

UK situation – Pre 2010 • Aim: How can we maximise the accuracy of evaluations?

– Using all existing data

– Without bias to existing evaluations

• Not all recorded data was being used – Some breeds excluded altogether

– Crossbreds largely excluded

– Not all breeds had full set of traits evaluated

• Not all data was being used optimally – Split proofs for the same bulls across breeds

– Suboptimal use of pedigree contributions

– Herdmate contemporaries not always included

• Growing interest in crossbreeding

Breed proportion - Changes

Breed proportions 2013 – Live cows

• 20% of cows not pure (>87.5% purity)

– Most are result of breed replacement

– 89% are >75% ‘pure’

• 5.3% are 1st generation crosses – Up 1.5% during last five years

Dealing with mixed breed data

• Correction for difference in variance

• Fitting full pedigree – Separate groups for unknown parents by breed

– Widespread use of AI has established many links

• Correction for Heterosis / Recombination – Crosses between four main breed groups considered

• Holstein

• British Friesian

• Reds (Ayrshire, Shorthorn, Brown Swiss, Montbeliarde)

• Other (Jersey, Guernsey, rest)

Example animal - 11779014

• Animal Breed Code %Breed Origin

• 11779014 68 50.00 NZ Jersey

• 11779014 76 12.50 N. American Jersey

• 11779014 04 6.25 UK Jersey

• 11779014 66 6.25 Danish Jersey

• 11779014 78 25.00 NZ Ayrshire

Heterosis of 5% • Offspring better than average of its parents

0

1000

2000

3000

4000

5000

6000

7000

8000

9000

Breed 1 Offspring Breed 2

Useful Heterosis • Offspring are better than either of their parents

0

1000

2000

3000

4000

5000

6000

7000

8000

9000

Breed 1 Offspring Breed 2

All breed evaluations- background

• Already routinely used in other countries: – E.g. Ireland, The Netherlands, New Zealand, and USA

• DairyCo commissioned feasibility study (‘07/08)

• Results of feasibility were promising – EGENES undertook further development work (08/09)

– International validation run in August 2009 (interbull)

– Implementation in January 2010

Impact

Largest changes for:

– Bulls used heavily in crossbreeding

– Bulls with limited information • Few daughters

• Few herds

• Therefore;

– Smaller breed populations relatively more change

• But also have largest gains in reliability

All-breed evaluations • Best use of all data; Two examples

• Morwick Sand Ranger (Red Holstein)

– Pure-bred analysis;

• 399 daughters in Holstein proof

• 392 daughters in Ayrshire proof

• 12 daughters in Shorthorn proof

– All-breed 837 dtrs in combined proof

• B Jurist (Swedish Red)

– Pure-bred analysis; • 0 dtrs in Holstein proof (not allowed)

• 127 dtrs in Ayrshire proof

• 21 dtrs in Shorthorn proof

– All-breed 766 dtrs in combined proof

Presentation of proofs

• Each animal receives only one proof • Post evaluation

– Animals get assigned to breed groups

– Each breed group has own genetic base • Reset in January 2010 to average of cows born in 2005

• Example:

• £PLI index applied to all breeds (Guernsey has own Merit Index)

Bull name Original PTA Re-based PTA

Milk SCC Base Milk SCC

Rosedale Advantage-Red -269 11 HOL -269 11

T-Bruno -279 -6 AYR 414 -4

Lakemead Rancher -275 7 FRI 406 13

On-going requirements

• Accurate data needed (lots of it !)

– Currently >100M records used

• Accurate animal identification

• Harmonised trait definitions (ICAR)

• Sharing (pooling) of Data

– Internationally

Conclusion and Future

• All-breed evaluations implemented in 2010 • Improved Accuracy of evaluations – within and across breeds

• New breed and trait evaluations added

• Industry response has been positive – Separate breed lists helped this situation

– However, one single list would help those x-breeding

• Future possibly All-breed genomic evaluations – Within breed genomics for Holstein - 2012

– More R&D needed to ‘translate’ DNA info to other breeds

top related