Last update: February 5, 2014 10:42:24 AM E-mail Print

 

ESTIMATION OF GENOTYPE X ENVIRONMENT INTERACTION FOR PRODUCTION TRAITS IN FINE AND STRONG WOOL MERINO SHEEP OF SOUTH AFRICA

 

W.J. Olivier1#, S.W.P. Cloete2,3, M.A. Snyman1 and J.B. van Wyk4

 

1Grootfontein Agricultural Development Institute, Private Bag X529, Middelburg (EC), 5900

2Department of Animal Sciences, University of Stellenbosch, Stellenbosch, 7602

3Institute for Animal Production: Elsenburg, Private Bag X1, Elsenburg, 7607

4Department of Animal, Wildlife and Grassland Sciences, University of the Free State, Bloemfontein, 9300

#E-mail: Willem Olivier

 

INTRODUCTION 

Genotypes are normally defined as different breeds, however in Merino sheep different strains or bloodlines are seen as different genotypes (McGuirk, 2009). These different strains are defined by differences in production and reproduction potential (Short & Carter, 1955; Dunlop, 1962; 1963; Jackson & Roberts, 1970; Mortimer et al., 1985; Mortimer & Atkins, 1989; Kleemann et al., 2006; McGuirk, 2009).

This perception that different genotypes may perform differently in the same environment or that the same genotypes may perform differently in different environments has led to the genotype by environment interaction (GxE) concept. There are different approaches to investigating GxE. When the same genotypes are used in different environments the most common method to evaluate this interaction is through the genetic correlation between the estimated breeding values in each environment (Falconer, 1952).

A GxE interaction can be defined as the change in the performance in a specific trait of different genotypes in different environments. This implies that the animal with the best genetic merit in a specific environment will not necessarily be the best performer or produce the best offspring in another environment, i.e. re-ranking of animals in different environments is expected (Falconer & Mackay, 1996; Lynch & Walsh, 1998; Cardoso & Tempelman, 2012).

This interaction may also be assessed by estimating the GxE variance as a variance ratio, expressed relative to the phenotypic variance (Dickerson, 1962). The phenotypes of animals across environments are therefore not always simply the sum of the genetic and environmental variation because it can also be affected by the interaction between the genotype and environment (Peaston & Whitelaw, 2006; Steinheim et al., 2008).

GxE interactions are considered to be present when the correlation between estimated breeding values for a specific trait expressed in different environments differs from unity (rG (genetic correlation) < 1; Falconer, 1952). Because it becomes cumbersome to derive correlations for all GxE combinations, there is a tendency to assume that the correlations between different environments for a trait affected by GxE are equal to unity and subsequently the GxE are excluded (Bertrand et al., 1985; Woolaston, 1987; Cameron & Curran, 1995; Warner et al., 2010). The exclusion of such interactions can compromise selection efficiency (Dominik & Kinghorn, 2008; Huquet et al., 2012).

The optimum use of specific genotypes in different environments depends on knowledge of the effect of GxE interactions on the important production traits (Vostrý et al., 2008; Warner et al., 2010). This interaction can potentially be important for wool sheep in South Africa, as production environments vary a lot from arid conditions with a low carrying capacity to high-potential irrigated pastures where high levels of production are sustained. It is also conceded that specific strains adapt better to specific environments/conditions than others (Dickerson, 1962).

The popular belief in South Africa is that fine wool cannot be produced successfully under the extensive and arid farming conditions. This contention resulted in a study where fine wool animals were evaluated on veld (Olivier & Olivier, 2007)

Therefore, the aim of this study was to estimate the GxE variance in a dataset that consisted of two genotypes that were farmed with in three different environments.

 

MATERIALS AND METHODS 

Data collected from 1989 to 1999 on the Grootfontein Merino flock (GMF), from 1989 to 2010 on the Cradock fine wool Merino stud (CMS) and from 1966 to 2010 on the Grootfontein Merino stud (GMS) were used for estimating the GxE interaction. The GMF was maintained at the Grootfontein Agricultural Development Institute (GADI) near Middelburg (31°28'S, 25°1'E) in the north-eastern Karoo region of South Africa. GADI is located in the False Upper Karoo (Acocks, 1988) and has an average annual rainfall of 375 mm.

During 1989, 400 Merino ewes of the Grootfontein Merino flock with an average fibre diameter of 23.6 μm were randomly divided into two groups of 200 each, subsequently labelled as a fine woolled (F) line and a control (C) line. The F-line was upgraded to produce finer wool by being mated to genetic fine wool rams from the Cradock fine wool Merino stud, while the C-line was mated to rams from the GMS.

The CMS are run on irrigated pastures at the Cradock Experimental Station near Cradock in the Eastern Cape province of South Africa. Olivier et al. (2006) gives a detailed description of the management and selection practices of this stud. The GMS was run on a combination of Karoo veld and irrigated pastures at GADI. Olivier (1989) and Olivier (1998) give a detailed description of the management and selection practices of this stud.

The production traits included in the analysis of the GxE interaction were weaning weight (WW), 15-month body weight (BW), greasy fleece weight (GFW; corrected to 365 days wool growth), clean fleece weight (CFW), fibre diameter (FD) and staple length (SL). The least-squares means (LSM) and the standard errors for these production traits were obtained with the PROC GLM-procedure of SAS and the significance levels of differences among the flocks were obtained with the PDIFF-option under the PROC GLM-procedure of SAS (SAS, 2009).

The fixed effects tested for significance for the GxE analysis included stud (CMS, GMS, fine wool line GMF (GMFF) and control line GMF (GMFC)), sex (males and females), year of birth, age of the dam (years), rearing status and age of the animal (linear regression) at weaning and 15 months of age.

The estimation of the genetic parameters was done with ASREML (Gilmour et al., 2009) by fitting single-trait animal models. GxE was estimated for the respective traits by adding sire x flockyear (combination of flock and year of birth) as an additional random effect to the operational model for each trait and the data were analysed with ASREML (Gilmour et al., 2009).

The investigation into the possible effect of GxE was done by estimation of the genetic correlation for the respective traits in the different environments. The genetic correlations were estimated using two-trait models with ASREML (Gilmour et al., 2009). The Spearman rank-order correlation (SAS, 2009) was used to estimate the correlations between the breeding values obtained from the models that include or exclude GxE.

 

RESULTS AND DISCUSSION 

The production data of the progeny from the three flocks are summarised in Table 1. The animals in the GMS and CMS were both heavier and produced more and longer wool than the two GMF lines. The two lines of the GMF and CMS produced finer wool than the GMS. Year of birth, sex, rearing status and age of the dam also affected the body weight and wool characteristics of the three studs.

 

Table 1. Least squares means depicting the effect of flock on weaning weight (WW), body weight (BW), greasy fleece weight (GFW), clean fleece weight (CFW), fibre diameter (FD) and staple length (SL) (± s.e.) of the three flocks

Trait

GMS

(n=13392)

CMS

(n=7770)

GMFF

(n=1996)

GMFC

(n=1896)

WW (kg)

26.7a ± 0.2

26.7a ± 0.2

21.1b ± 0.3

21.9c ± 0.3

BW (kg)

51.0a ± 0.5

62.2b ± 0.5

44.2c ± 1.1

47.4c ± 1.3

GFW (kg)

11.4a ± 0.3

11.3a ± 0.3

8.3b ± 0.3

8.7b ± 0.3

CFW (kg)

5.0a ± 0.1

5.3a ± 0.1

3.1b ± 0.1

3.3b ± 0.1

FD (μm)

21.1a ± 0.1

19.5b ± 0.1

18.3c ± 0.1

19.6b ± 0.1

SL (mm)

97.4a ± 0.7

100.6b ± 0.7

79.3c ± 0.9

76.6d ± 0.9

a,b,c - Values with the different superscripts differed significantly (P<0.05); GMS – Grootfontein Merino stud; CMS – Cradock fine wool Merino stud; GMFF – fine wool line of the Grootfontein Merino flock; GMFC – control line of the Grootfontein Merino flock

 

The genetic correlations between the traits in different environments are summarised in Table 2. The genetic correlation for WW between GMS and GMF and SL between GMS and GMF were not different from unity through association with the standard errors. A GxE interaction is of agricultural importance when the genetic correlation for a trait between different environments is below 0.80 (Robertson, 1959). The genetic correlations, through association with the standard errors for WW (GMSxGMF), FD and SL (CMSxGMF) were higher than 0.80. This suggests that there is no GxE present for these traits. The genetic correlations for BW and fleece weight between the different environments indicate that there are a GxE present.

 

Table 2. Genetic correlations ( s.e.) between the same trait in different environments on weaning weight (WW), body weight (BW), greasy fleece weight (GFW), clean fleece weight (CFW), fibre diameter (FD) and staple length (SL)

 

WW

BW

GFW

CFW

FD

SL

GMS x GMF2

0.73 ± 0.19

0.55 ± 0.21

0.54 ± 0.17

0.42 ± 0.30

0.74 ± 0.14

0.98 ± 0.08

CMS x GMF2

0.95 ± 0.10

0.40 ± 0.23

0.52 ± 0.19

0.42 ± 0.32

0.80 ± 0.12

0.80 ± 0.13

2 GMS x GMF = Genetic correlation between Grootfontein Merino stud (GMS) and Grootfontein Merino flock (GMF); CMS x GMF = Genetic correlation between Cradock Merino stud (GMS) and Grootfontein Merino flock (GMF)

 

The correlations obtained in the current study for BW and fleece weights are lower than the value reported by Dominik et al. (1999) but higher than the value reported by MacLeod et al (1990). For FD and SL the estimates of the current study are in agreement with the values estimated by Dominik et al. (1999). MacLeod et al. (1990) however, reported a much lower estimate for FD than that of the current study and Dominik et al. (1999).

The variance ratios for WW, BW, GFW, CFW, FD and SL are summarised in Table 3. The ge2 (GxE variance ratio) effect for WW, BW, GFW, CFW, FD and SL amounted to 0.01 ± 0.00, 0.08 ± 0.01, 0.10 ± 0.01, 0.05 ± 0.01, 0.02 ± 0.00 and 0.03 ± 0.01 respectively. The estimates for WW and BW are in agreement with the values reported in the literature (Hagger, 1998; Maniatis & Pollott, 2002; Pollott & Greeff, 2004). The ge2 values estimated in this study for CFW, FD and SL are in accordance with values in the literature, while the current estimate for GFW is slightly higher than the reported value (Pollott & Greeff, 2004).

 

Table 3. Variance ratios (± s.e.) for weaning weight (WW), body weight (BW), greasy fleece weight (GFW), clean fleece weight (CFW), fibre diameter (FD) and staple length (SL)

Variance ratio

WW

BW

GFW

CFW

FD

SL

h2a

0.14 ± 0.02

0.37 ± 0.05

0.38 ± 0.02

0.33 ± 0.02

0.55 ± 0.02

0.33 ± 0.02

h2m

0.06 ± 0.01

0.05 ± 0.01

0.04 ± 0.01

0.04 ± 0.01

 

 

ram

-0.19 ± 0.01

 

 

 

 

 

c2mpe

0.09 ± 0.01

 

 

 

 

 

ge2

0.01 ± 0.00

0.08 ± 0.01

0.10 ± 0.01

0.05 ± 0.01

0.02 ± 0.00

0.03 ± 0.00

h2a = direct additive heritability, h2m = maternal heritability, ram = genetic correlation between the animal effects, c2mpe = maternal permanent environmental effect, ge2 = GxE variance ratio

 

The Spearman rank-order correlations between the estimated breeding values obtained for models including or excluding the sire x flockyear variance are summarised in Table 4. It is evident from this table that the inclusion of a sire by flockyear variance component will not lead to a large scale re-ranking of the animals in this analysis.

In summary, the estimates obtained in the current study, as well as in the literature indicated that the magnitude of ge2 for body weight and fleece traits of sheep were generally below 0.10. Furthermore, the genetic correlations among most of the traits expressed in different environments were above 0.80 which implies that GxE for that trait was not of agricultural importance. The Spearman ranking correlations support the above mentioned results, as these values were not indicative of large scale re-ranking of animals. However, across-flock evaluation of production traits in the Merinoplan (South African performance testing scheme for Merinos) may be subject to GxE of larger magnitude.

 

Table 4. Spearman rank-order correlations between the ranks obtained from models including or excluding sire x flockyear variance for the different traits

Trait

Spearman rank correlation

Weaning weight (WW)

0.997

Body weight (BW)

0.985

Greasy fleece weight (GFW)

0.983

Clean fleece weight (CFW)

0.989

Fibre diameter (FD)

0.998

Staple length (SL)

0.995

 

 

CONCLUSION

It is evident from the results of this study that BW and GFW are affected at a higher level of GxE interaction compared to the other traits. Furthermore, these small GxE interactions did not led to a large-scale re-ranking of sires in different production environments.

However, it is important that the possibility of GxE interactions are considered when breeding values is estimated on a national basis where different genotypes and environments are included in the same analysis. Ignoring GxE may lead to the estimation of biased breeding values which can have a detrimental effect on the selection of replacement animals.

 

REFERENCES

Acocks, J.P.H., 1988. Veld Types of South Africa. 3rd ed. Botanical Research Institute, Dept. of Agric. and Water Supply, South Africa.

Bertrand, J.K., Berger, P.J. & Willham, R.L., 1985. Sire x environment interactions in beef cattle weaning weight field data. J. Anim. Sci. 60, 1396–1402.

Brash, L.D., Taylor, P.J. & Gilmour, A.R., 1997. Estimates of genetic parameters and environmental effects for production traits in young Merino rams. Proc. Proc. Assoc. Advmt. Anim. Breed. Genet. 12, 529-533.

Cameron, N.D. & Curran, M.K., 1995. Genotype with feeding regime interaction in pigs divergently selected for components of efficient lean growth rate. Anim. Sci. 61, 123–132.

Cardoso, F.F. & Tempelman, R.J., 2012. Linear reaction norm models for genetic merit prediction of Angus cattle under genotype by environment interaction. J. Anim. Sci. 90, 2130-2141.

Dickerson, G.E., 1962. Implications in genetic-environmental interaction in animal breeding. Anim. Prod. 4, 47-63.

Dominik, S., Crook, B.J. & Kinghorn, B.P., 1999. Genotype x management interaction on wool production traits and body weight in Western Australian merino sheep. Proc. Assoc. Advmnt. Anim. Breed. Genet. 13, 98-101.

Dominik, S. & Kinghorn, B.P., 2008. Neglecting genotype x environment interaction results in biased predictions from selection index calculations. Livest. Sci. 114, 233-240.

Dunlop, A.A., 1962. Interactions between heredity and environment in the Australian Merino. I. Strain X location interactions in wool traits. Aust. J. Agric. Res. 13, 503-531.

Dunlop, A.A., 1963. Interactions between heredity and environment in the Australian Merino. 11. Strain X location interactions in body traits and reproductive performance. Aust. J. Agric. Res. 14, 690-703.

Falconer, D.S., 1952. The problem of environment and selection. Am. Nat. 86, 293–298.

Falconer, D.S. & Mackay, T.F.C., 1996. Introduction to quantitative genetics. 4th ed Longman Group Ltd. Harlow, UK.

Gilmour, A.R., Gogel, B.J., Cullis, B.R. & Thompson, R., 2009. ASReml User Guide Release 3.0 VSN International Ltd, Hemel Hempstead, HPI 1ES, UK.

Hagger, C. 1998. Litter, permanent environmental, ram-flock and genetic effects on early weight gain of lambs. J. Anim. Sci. 76, 452–457.

Hill, J.A., 2001. Phenotypic and genetic parameters for the S.A. strongwool Merino strain with an emphasis on skin characters as early indicators of wool productivity. PhD Thesis, University of Adelaide, Adelaide.

Huquet, B., Leclerc, H. & Ducrocq, V., 2012. Modelling and estimation of genotype by environment interactions for production traits in French dairy cattle. Gen. Sel. Evol. 44, 35-48.

Jackson, N. & Roberts, E.M., 1970. Comparison of three Australian Merino strains for wool and body traits. I. Genetic means of studs and strains and their interactions with years and sexes. Aust. J. Agric. Res. 21, 815-835.

Kleemann, D.O., Grosser, T.I. & Walker, S.K., 2006. Fertility in South Australian commercial Merino flocks: aspects of management. Therio. 65, 1649-1665.

Lynch, M. & Walsh, B., 1998. Genetics and analysis of quantitative traits. 1st Ed. Sinauer Associates, Inc. Sunderland, MA.

MacLeod, I.M., Howe, R.R., Lewer, R.P. & Woolaston, R.R., 1990. Genotype by environment interactions among Merino sheep in Western Australia. Proc. Aust. Assoc. Anim. Breed Genet. 8, 303-306.

Maniatis, N. & Pollott, G.E., 2002. Genotype x environment interactions in lamb weight and composition traits. Anim. Sci. 75, 3–14.

McGuirk, B.J., 2009.  Genotype x environment interactions and Merino breeding programmes for wool production. Proc. Assoc. Advmnt. Anim. Breed. Genet. 18, 42-47.

Mortimer, S.I. & Atkins, K.D., 1989.  Genetic Evaluation of Production Traits between and within Flocks of Merino Sheep. I Hogget Fleece Weights, Body Weight and Wool Quality. Aust. J. Agric. Res. 40, 433-443.

Mortimer, S.I., Atkins, K.D. & Rogan I.M., 1985. Reproductive performance of Merino strains and bloodlines. Proc. Aust. Assoc. Anim. Breed Genet. 5, 218–219.

Olivier, J.J., 1989. Genetic and environmental trends in the Grootfontein Merino stud. Ph.D. thesis, University of the Orange Free State, Bloemfontein, South Africa.

Olivier, W.J., 1998. Evaluation of ewe productivity measures.  M.Sc. Agric. Thesis, University of the Orange Free State, Bloemfontein, South Africa.

Olivier, W.J. & Olivier, J.J., 2007. Evaluation of genetic fine wool animals under natural conditions in the non-traditional fine wool producing areas of the areas. Grootfontein Agric 7, 35-38.

Olivier, W.J., Olivier, J.J., Cloete, S.W.P. & Van Wyk, J.B., 2006. Genetic analysis of the Cradock fine wool Merino stud. Proc. 8th World Congr. Genet. Appl. Livest. Prod., Belo Horizonte, 13-18 Augustus, 599-602.

Peaston, A. & Whitelaw, E., 2006. Epigenetics and phenotypic variation in mammals. Mammalian Genome, 17, 365−374.

Pollott, G.E. & Greeff, J.C., 2004. Genotype x environment interactions and genetic parameters for fecal egg count and production traits of Merino sheep. J. Anim. Sci. 82, 2840–2851.

Rao, S. & Notter, D.R., 2000. Genetic analysis of litter size in Targhee, Suffolk, and Polypay sheep. J. Anim. Sci. 78, 2113-2120.

Robertson, A., 1959. The sampling variance of the genetic correlation coefficient. Biomet., 15, 469-485.

SAS, 2009. SAS Online Doc, Version 9.2. SAS Institute Inc. Cary, N.C., USA.

Short, B.F. & Carter, H.B., 1955. Analysis of Merino stud flocks. CSIRO Aust. Bull. No. 276.

Steinheim, G., Ødegård, J., Ådnøy, T. & Klemetsdal, G., 2008. Genotype by environment interaction for lamb weaning weight in two Norwegian sheep breeds. J. Anim. Sci. 86, 33-39.

Vostrý, L., J. Přibyl, J., Jakubec, V., Vesela, Z. & Majzlik, I., 2008. Selection of a suitable definition of environment for the estimation of genotype x environment interaction in the weaning weight of beef cattle. Czech J. Anim. Sci., 53, 407-417.

Warner, R.D., Greenwood, P.L., Pethick, D.W. & Ferguson, D.M., 2010. Genetic and environmental effects on meat quality. Meat Sci. 86, 171-183.

Woolaston, R.R., 1987. Genotype x environment interactions and their possible impact on breeding programs. In: Merino Improvement Programs in Australia, (ed. BJ McGuirk), Australian Wool Corporation, Melbourne, pp. 421-435.

 

Published

Grootfontein Agric 14 (1) (25)