Last update: December 1, 2010 04:08:20 PM E-mail Print

 

SHORT COMMUNICATION

HERITABILITY ESTIMATES FOR MERINO SHEEP OBTAINED

FROM A NATIONAL PROGENY TEST

 

P.G.J. Groenewald * , J.J.Olivier

Agricultural Research Council, Middelburg Regional Office, Private Bag X529,

 Middelburg EC, 5900 Republic of South Africa

 

W.J. Olivier

Grootfontein Agricultural Development Institute,  Private Bag X529,

Middelburg EC, 5900 Republic of South Africa

 

Telephone  : 027 4924 22563

Fax             : 027 4924 21343

 * Author to whom correspondence should be addressed

 


Data collected through the National Progeny Test for Merino sheep from 1991 to 1996 were used for this study.  The data included 5342 progeny from 232 sires in 30 genetically linked flocks. The heritabilities for  several production - and type traits were estimated by means of Hendersons method III, using the LSMLMW programs of Harvey.  The following traits were analysed: body weight (BW) , clean fleece weight (CFW), mean fibre diameter (MFD), staple length (SL), pleat score (PS), wool quality (QUAL), variation in crimps over the fleece (VAR), wool yolk (OIL), staple formation  (SF), belly and points (BP), head size  (HEAD), colour (COL), fore quarters (FQ), pasterns (PAS), hocks  (HOC) and conformation (CON).    In most traits the effect of herd-year-season  (HYS) was significant (P # 0.01).  For the production traits, excluding MFD, the interaction between HYS and sex  was significant.  Birth status had no significant effect on QUAL, OIL, SF, COL and PAS. Heritability estimates for the various traits were: 0.34 for BW, 0.23 for CFW, 0.44 for MFD, 0.70 for SL, 0.32 for PS, 0.27 for QUAL, 0.23 for VAR, 0.24 for OIL, 0.09 for SF, 0.17 for BP, 0.35 for HEAD, 0.17 for COL, 0.21 for FQ, 0.12 for PAS, 0.26 for HOC and 0.31 for CON.  Standard errors for these estimates ranged from  0.03 to 0.06. It is recommended that these heritability estimates, calculated from industry data,  be used in the construction of selection index weighting factors, as well as EPD=s in the National Progeny Testing Scheme.

 

Data wat van 1991 tot 1996 d.m.v. die Nasionale Merino Nageslagtoets ingesamel is, is vir die studie gebruik.  Die data bestaan uit  5342 nageslag van 232 vaders in 30 geneties gekoppelde-kuddes. Oorerflikhede vir verskeie produksie - en subjektiewe eienskappe is deur middel van Henderson se metode III beraam deur van die LSMLMW programme van Harvey gebruik te maak.  Die volgende eienskappe is ontleed: liggaamsgewig (BW) , skoonvaggewig (CFW), gemiddelde veseldikte (MFD), stapellengte (SL), plooitelling (PS), wolkwaliteit (QUAL),  variasie in kartels oor die vag  (VAR), wololie (OIL), stapelformasie (SF), onderlyne (BP), kopgrootte (HEAD), kleur (COL), voorkwarte (FQ), kote (PAS), hakke (HOC) en algemene bouvorm (CON).  Die effek van kudde-jaar-seisoen (HYS) was betekenisvol (P # 0.01) vir die meeste eienskappe.  Die interaksie van HYS met geslag was betekenisvol vir al die produksie-eienskappe, behalwe MFD.  Geboortestatus  het geen betekenisvolle effek op QUAL, OIL, SF, COL en PAS gehad nie. Die oorerflikhede vir die verskillende eienskappe was soos volg: 0.34 vir BW, 0.23 vir CFW, 0.44 vir MFD, 0.70 vir SL, 0.32 vir PS, 0.27 vir QUAL, 0.23 vir VAR, 0.24 vir OIL, 0.09 vir SF, 0.17 vir BP, 0.35 vir HEAD, 0.17 vir COL, 0.21 vir FQ, 0.12 vir PAS, 0.26 vir HOC en 0.31 vir CON.  Die standaardfoute vir hierdie beramings het van 0.03 tot 0.06 gewissel.  Dit word aanbeveel dat die oorerflikhede, beraam vanaf bedryfsdata,  gebruik word in die opstel van seleksie-indeks wegingsfaktore, sowel as EPD=s in die Nasionale Nageslagtoetsskema.

 

A selection programme for Merino sheep  was suggested by Poggenpoel & van der Merwe (1975) and it has since been widely implemented in the stud industry. One of the shortcomings of this programme was that variances and covariances for production traits used for the construction of a selection index were estimated from results in the literature.  Furthermore, in the past, all heritability estimates for South African Merino sheep were obtained from closed experimental flocks and these may differ from industry-based data.

 

The National  Progeny testing scheme of the Merino Breeder=s Society forms  an  integral part of the South African Small Stock Performance- and Progeny Testing Scheme.   The purpose of the National Progeny test is to identify superior AI-rams and to supply the Merino industry with the vital genetic links between flocks, in order to calculate breeding values for the Merino breed on a national basis.  

 

In the National Progeny test several production traits, as well as 11 type traits  were recorded.  Type traits were recently introduced into the National Progeny test and therefore no heritability estimates were available.  The objective of  this study was therefore to estimate  heritabilities of production traits, as well as type traits, using industry based data to compliment the parameters used in the national based breeding plans.

 

Data collected in 30 genetically linked flocks over a six year period  (1991 - 1996) were used for the analysis.  The data were collected  through the National Progeny Testing Scheme of the Merino Stud Breeder=s Society, in collaboration with the Animal Improvement Institute of the Agricultural Research Council.  At 15 to 18 months of age all animals were performance tested as described by Erasmus et al. (1990).    Production traits included in the analysis were body weight (BW), clean fleece  weight (CFW), mean fibre diameter (MFD), pleat score (PS) and staple length (SL).    

 

Scoring of type traits, on a linear scale,  was performed approximately one month before shearing as described by Olivier et al. (1987).  Type traits included in the analyses were wool quality (QUAL),  variation in crimps over the fleece (VAR), wool yolk (OIL), staple formation (SF), belly and points (BP), size of head (HEAD), colour (COL),  fore quarters (FQ), pasterns (PAS), hocks (HOC) and conformation (CON). 

 

Only sires with more than five progeny were retained for analysis which resulted in the exclusion of  14 sires and their progeny. After editing, data were available for 5342 progeny (2530 ram and 2812 ewe progeny) from 232 sires, giving an effective number of  20,65 progeny per sire. 

 

In an analyses of variance, using mixed model least-squares procedures (Harvey, 1990), the following general mixed model was fitted for each trait:

 

Yijklmn = μ + pi + bj + sk + HYSl + adm + (HYSb)lj + (HYSs)lk + eijklmn

Where:

Y ijklmn = observation of the j th birth status of the k th sex of the l th herd-year-season of the mth age                of dam on the nth progeny for sire i,

μ = population mean of the trait,

 pi   = random effect of the ith sire (i = 1, ... 232),

 bj = effect of the jth birth status (j = 1,2),

sk = effect of the kth sex (k = 1,2),

 

HYSl = effect of the lth herd-year-season (l = 1,... 36),

adm = effect of the mth age of dam (m = 1,2),

(HYSb)lj = effect of the interaction between the lth herd-year-season and the jth birth status,

(HYSs) lk = effect of the interaction between the lth herd-year-season and the kth sex and

eijklmn   = random error

*****

The identities of the dams were not known.  According to Erasmus et al.  (1990) parameter estimates obtained from sire models using REML or Henderson=s method III are in close agreement  and there seems little point in using REML unless all relationships among animals can be utilized.   Therefore heritabilities were estimated by means of Henderson=s method III using the LSMLMW programmes of Harvey (1990).  Only  significant fixed effects for each trait were included in the model. The following model was fitted for each trait to estimate heritabilities:

 

y = Xb + Zp + e

where

y is a vector of observed traits of animals;

b is a vector of fixed effects;

p is a vector of random sire effects;

X and Z are the corresponding incidence matrices relating to the effects to y and

e is the vector of residuals

 

The means, standard deviations and coefficients of variation (CV) for all traits analysed, are given  in Table 1.    The means of the production traits  are in the same order as reported by Poggenpoel & van der Merwe (1992) and Snyman et al. (1996), but lower than those calculated  by Olivier et al. (1997).  These differences are  due to environmental differences.

 

The CV of MFD, CFW and BW accords well  with results in the literature (Cloete et al., 1992; Olivier et al., 1997).   The CV of type traits ranges from 10 to 40 % and corresponds with results from  Olivier et al. (1997).  The relatively lower CV for OIL and HOCKS may be due to the linear scoring method of these traits.  For these traits the intermediate value was considered as ideal.   In general the means of the type traits accords with the results of Olivier et al.  (1997).

From Table 2 it is evident that in all traits analysed the effect of herd-year-season (HYS) was significant (P# 0.01).  For the production traits, excluding MFD, the interaction between HYS and sex was significant.  Birth status had no significant effect (P$0.05) on QUAL, OIL, SF, COL and PAS.  The significant effect of sex for type traits could probably be ascribed to the different  environments under which rams and ewes were raised.  Age of dam  proved to be non significant for all traits analysed  and were therefore excluded.  

 

Heritability estimates for BW and CFW (Table 3) are slightly lower than heritabilities reported by Poggenpoel & van der Merwe  (1975), Van Wyk et al. (1985) and Olivier et al. (1997). Most other heritability estimates obtained in this study were in the same order as heritability estimates calculated by  Olivier et al. (1997) on research flocks.   Heritability estimates for MFD and PS accords well with those reported for South African Merino sheep, but the estimate for SL was higher than heritability estimates  reported in literature (Van Wyk et al., 1985; Olivier et al.,1997) and should be viewed with caution.   

 

The breeding objective of the Merino industry  is to increase body weight, keep wool weight constant and to reduce mean fibre diameter and number of pleats.  Most breeders achieve this by applying  the selection index theory as described by Poggenpoel & van der Merwe (1975).   The heritability estimates used by Poggenpoel & van der Merwe (1975) to calculate these selection index weighting factors differ slightly from the heritability estimates in  this study.  Therefore it can be recommended that these industry based heritability estimates be used for construction of national as well as individual selection indices.  The estimates calculated for type traits can also be implemented in the National Progeny Testing Scheme for the calculation of EPD=s.

 

 

References

CLOETE, S.W.P., DELPORT, G.J., ERASMUS, G.J., OLIVIER, J.J., HEYDENRYCH, H.J. & DU TOIT, ELIZABETH, 1992.  Environmental and genetic trends in clean fleece mass,  live mass and fibre diameter in selection and control flocks involving a selection experiment for increased clean fleece mass in South African Merino sheep.S. Afr. J. Anim. Sci. 22 (2), 50 - 57.

ERASMUS, G.J., DE LANGE, A.O., DELPORT, G.J. & OLIVIER, J.J., 1990.  Genetic and phenotypic  parameter estimates of production traits in Merino sheep in an arid environment.  S. Afr. J. Anim. Sci. 20, 31 - 34.

HARVEY, W.R., 1990.  User=s guide for LSMLMW and MIXMDL (PC-2 version). Mixed model least squares and maximum likelihood computer program. Ohio state Univ., Ohio, USA.

OLIVIER, J.J., DELPORT, G.J., ERASMUS G.J. & EKSTEEN, T.J., 1987.  Linear type scoring in Merino sheep.  Karoo Agric., 3(9),1-4.

OLIVIER, J.J., CLOETE, S.W.P. & M.A. SNYMAN, 1997.  Relationship between type and production traits in South African Merino sheep.  Proc. South African Soc of Anim. Sci. 35th Congress, Nelspruit.

POGGENPOEL, D.G. & VAN DER MERWE, C.A., 1975.  Die gebruik van seleksie-indekse by Merinoskape. S. Afr. J. Anim. Sci. 5, 249-255.

POGGENPOEL, D.G. & VAN DER MERWE, C.A., 1992.  Between-flock genetic differences in 40 Merino studs. S. Afr. J. Anim. Sci. 22(6), 175-180.

SNYMAN, M.A., OLIVIER, J.J. & OLIVIER, W.J.,1996.  Variance components and genetic parameters for body weight and fleece traits of Merino sheep in an arid environment. S. Afr. J. Anim. Sci. 26 (1), 11-14.

VAN WYK, J.B., NEL, J.A., VAN DER SCHYFF,W. & ERASMUS, G.J., 1985.  Fenotipiese - en genetiese parameters by >n kommersiële kudde Merinoskape. S. Afr. J. Anim. Sci.15 (4), 171-172.

 

 

Table 1:   Means, standard deviations and coefficient of variation of  production and type traits

Trait

Mean

SD

CV (%)

BW

36.66 (kg)

7.18

19.59

CFW

3.43  (kg)

0.98

28.57

MFD

19.31 (μ)

1.33

6.89

SL

10.25 (cm)

1.62

15.80

PS

4.70

1.96

41.70

QUAL

29.44

7.28

24.73

VAR

32.35

7.95

24.57

OIL

24.58

2.56

10.41

SF

28.36

6.44

22.71

BP

25.15

6.34

25.21

HEAD

27.96

6.52

23.32

COL

35.03

9.51

27.15

FQ

26.29

5.66

21.53

PAS

36.65

7.57

20.65

HOC

23.65

4.10

17.34

CON

27.53

6.78

24.63

 

Table 2: Model specification for production  and type traits

EFFECT

HYS

HYS_s

B

s

df

TRAIT

35

 

67

 

1

 

1

 

BW

 

*

*

 

CFW

 

*

*

 

MFD

 

*

*

 

PS

*

 

*

*

SL

 

*

*

 

QUAL

*

 

 

*

VAR

*

 

*

*

OIL

*

 

 

*

SF

*

 

 

*

BP

*

 

*

*

HEAD

*

 

*

*

COL

*

 

 

*

FQ

*

 

*

*

PAS

*

 

 

*

HOC

*

 

*

*

CON

*

 

*

*

* = P # 0.01

df         -   degrees of freedom

HYS    -   interlaced effect of herd, year and season

HYS_s -  interlaced effect of herd, year, season and sex

B          -   birth status

s           -   sex

 

 

Table 3: Heritability estimates for production and type traits

Trait

Heritability

 

SE

BW

0.34

0.04

CFW

0.23

0.04

MFD

0.44

0.05

SL

0.70

0.06

PS

0.32

0.04

QUAL

0.27

0.04

VAR

0.23

0.04

OIL

0.24

0.04

SF

0.09

0.03

BP

0.17

0.03

HEAD

0.35

0.04

COL

0.17

0.03

FQ

0.21

0.03

PAS

0.12

0.02

HOC

0.26

0.04

CON

0.31

0.04