using single-step genomic blup (ssgblup) in american angus ·...

Daniela Lourenco, University of Georgia June 11, 2015

2015 BIF Gene>c Predic>on CommiCee 1

Genetic evaluation using single-step genomic BLUP

(ssGBLUP) in American Angus

Daniela Lourenco Ignacy Misztal, Keith Bertrand, Tonya Amen, Dan Moser

University of Georgia and Angus Genetics Inc. BIF 6/11/2015

Genomics

Genomics ???

G

Genomics

Data

Pedigree

GGenomic Info

Henderson, 1963

BLUP

ssGBLUP

H

single-step genomic BLUP

ssGBLUP - UGA

H

Lourenco et al., 2015 Aguilar et al., 2010

H-1 = A-1 +0 00 G-1 -A22

-1

⎡

⎣⎢⎢

⎤

⎦⎥⎥

ssGBLUP

ssGBLUP x Multistep

Multistep ssGBLUP



ssGBLUP x Multistep

Multistep ssGBLUP

•  Same models as in BLUP •  Raw phenotypes

•  Animals •  Genotyped •  Ungenotyped

•  Simple models

•  Pseudo-phenotypes

•  Animals •  Genotyped

Number of genotyped

> 8 million ????

http://www.angus.org/

Which animals should be genotyped?

AAA 82,000

•  1.7k bulls acc ≥ 0.85

•  300 cows ≥ 0.85

•  31k acc < 0.85

•  18k for validation

•  30k no phenotypes


> http://www.cattle.com/articles/title/Angus+Cattle.aspx

http://www.cattle.com/articles/title/Angus+Cattle.aspx

http://www.streamcattle.com/images/2008/oct20/Lot6.jpg




Top bulls Acc ≥ 0.85 N ~ 1,700

+ http://www.cattle.com/articles/title/Angus+Cattle.aspx

http://www.newmanangus.com/donors/SAVBlackcapMay4568.html

Top B+C Acc ≥ 0.85 N ~ 2,000

+ http://www.cattle.com/articles/title/Angus+Cattle.aspx


http://benfield-whiteridge.com/phil.htm http://www.hansenangusranch.com/

https://farmcentrefrancais.wordpress.com/page/2/

+

All N ~33,000

Validation N ~ 18,000


Validation Ability to predict future phenotypes for 18,000

“young” genotyped animals


1)  Remove phenotypes

2)  Run evaluations EBV GEBV


Correlation



8.2M

18,000 + 1,700 or + 2,000 or + 33,000

BW=6.2M

WW=6.9M

PWG=3.4M

Growth CE=1.3M

BW=6.2M

Calving ease



0.29

0.34 0.34

0.39

BLUP top bulls top B+C All

Pred

ic>ve ability

BW

+5 0 +5


0.29

0.34 0.34

0.39

0.34 0.35 0.35

0.38


Pred

ic>ve ability

BW WW

+1 0 +3

+5 0 +5


0.29

0.34 0.34

0.39

0.34 0.35 0.35

0.38

0.23

0.27 0.27 0.29


Pred

ic>ve ability

BW WW PWG

+4 0

+2

+1 0 +3

+5 0 +5


0.29

0.34 0.34

0.39

0.34 0.35 0.35

0.38

0.23

0.27 0.27 0.29

0.12 0.13 0.13 0.13


Pred

ic>ve ability

BW WW PWG CE

+4 0

+2

+1 0 +3

+5 0 +5

+1 0 0

Lots of genotyped animals

82,000

110,000

130,000



Lots of genotyped animals

Association/Species # Genotyped

Fish 2,500

Pigs 15,000

Chicken 18,000

AAA 130,000

US Holstein 850,000

Aguilar et al., 2010

H-1 = A-1 +0 00 G-1 -A22

-1

⎡

⎣⎢⎢

⎤

⎦⎥⎥

X Limitation ~150,000

Base and non-base animals

BASE NON-BASE






http://www.rieksblackangus.com/

Genomic recursions for G-1

G-1

BASE invert

NON-BASE


0.94

0.97

0.99 0.99 0.99 0.99

2k 4k 8k 10k 20k 33k

Correla>

on (G

EBV,GE

BV_recursio

n)

Defini>on of Base

WW -‐ Angus


0.94

0.97

0.99 0.99 0.99

2k 5k 10k 15k 20k

Correla>

on (G

EBV, GEB

V_recursion)

Defini>on of Base

US Holsteins -‐ Final Score


BASE NON-BASE






http://www.rieksblackangus.com/

BA

SE N

ON

-BA

SE

BASE NON-BASE



Indirect prediction for young

•  Young animals

•  Genotyped

•  No phenotypes

AAA 82,000

Interim evaluations for young animals


ALL ALL ALL

ssGBLUP direct prediction

GEBV = w1PA + w2YD + w3PC + w4DGV – w5PP

H-1 = A-1 +0 00 G-1 -A22

-1

⎡

⎣⎢⎢

⎤

⎦⎥⎥

yield deviation

progeny contribution


ALL ALL ALL

ssGBLUP direct prediction

ALL ALL Reference Animals

SNP effect

DGV for young = Zu GEBV ≈ w1PA + w2DGV – w3PP

Wang et al., 2012


INDIRECT predictions

GEBV ≈ w1PA + w2DGV – w3PP

DIRECT predictions

GEBV = w1PA + w2YD + w3PC + w4DGV – w5PP

VS


INDIRECT predictions

DIRECT predictions

H-1 = A-1 +0 00 G-1 -A22

-1

⎡

⎣⎢⎢

⎤

⎦⎥⎥

H-1 = A-1 +0 00 G-1 -A22

-1

⎡

⎣⎢⎢

⎤

⎦⎥⎥

VS 2,000 8,000

33,000

2,000 + 18k 8,000 + 18k

33,000 + 18k


0.23

0.19

0.24

0.27

PA DGV GEBV_ind GEBV_dir

Pred

ic>ve Ab

ility

ref_2k




0.23

0.19

0.24

0.27

0.23 0.24

0.26 0.27


Pred

ic>ve Ab

ility

ref_2k ref_8k


0.23

0.19

0.24

0.27

0.23 0.24

0.26 0.27

0.23

0.29 0.29 0.29


Pred

ic>ve Ab

ility

ref_2k ref_8k ref_33k

�  ssGBLUP

�  Modified BLUP

�  Genomic is added to improve evaluations

�  Easy to implement – modification in relationships

�  Ready for thousands – millions of genotyped

�  Flexible for interim evaluation of young animals

�  Suitable for genetic evaluation of Beef Cattle

Summary Acknowledgements

Yutaka Masuda

Ignacio Aguilar

Andres Legarra

Shogo Tsuruta

Breno Fragomeni

using single-step genomic blup (ssgblup) in american angus ·...

Documents