Predicting genetic change in staple strength : how much gain can we expect?

Abstract

Proc. Assoc. Advmt. Anim. Breed. Genet. Vol13 PREDICTING GENETIC CHANGE IN STAPLE STRENGTH - HOW MIJCII GAIN CAN WE EXPECT? J. A. HilI ' and R. W. Ponzoni ' . ' Department of Animal Science, The University of Adelaide, Waite, Glen Gsmond, SA 5064 ' South Australian Research and'Development Institute, GPO Box 397, Adelaide, SA 5001 SUMMARY Genetic change in staple strength (SS) was predicted using phenotypic and genetic parameters and approaches from two independent studies [South Australian (SAP&g) and Western Australian (WAp&g)]. The SAP&g treat trait expressions in the two sexes and at different ages (hogget or adult) as different traits, whereas the WAp&g treat them as a single trait. Clean fleece weight (CFW), fibre diameter (FD), coeftlcienti of variation of fibre diameter (CV) and, SS were `inclu&%l'in the breeding objective, and a range of selection indices and economic values for FD, CV and SS were investigated. Based on the more elaborate model provided by the SAP&g it was concluded that although genetic gain in SS was possible, it would be smaller and harder to'adhieve than earlier suggested by Western Australian studies. Keywords : Staple strength, genetic change, breeding objective, Merino sheep INTRODUCTION To date the message to ram breeders and woolgrowers regarding the ,prospects of genetic improvement of staple strength (SS) in Merino sheep has been mainly based on information derived from Western Australian studies (Lewer and Li 1994; Greeff et al. 1995; Greeff et al. 1997). The message has been one of high expectations with important gains in SS achieved relatively easily using the coefftcient of variation of fibre diameter (CV) as a selection criterion., Here we conduct a selection index study to predict and compare the genetic change in staple strength when CV and SS are included in the .breeding objective and in the selection index. This is done using South Australian parameters (SAP&g) as well as those derived from the Western program (WAp&g) (Greeff 1997 and Greeff pers. comm.). MATERIALS AND METHODS The data used in the estimation of the SAP&g were Merino Resource Flock (Ponzoni et al. 1995). The the ewe records were taken at 16,28 and 40 months of wool growth, respectively. Wool samples wer measurement of the wool characters (Table 1). Australian from the ram and ewe progeny of the Turretfield ram records were taken at 16 months of age and of age. Therams and ewes had.6 and 12 months e taken from the mid-side of .each fleece for Heritabilities and phenotypic and genetic correlations were estimated using ASREML (Gilmour et al. 1998). An animal model was fitted, including the fixed effects of year, stud, age of dam, type of birth and rearing class (ram and ewe data) and lambing and rearing status (ewe data only). Day of birth was fitted as a linear covariate. Parameters for 28 and 40 month old ewes were averaged to 46 Proc. Assoc. Advmt. Anim. Breed. Genet. Vol13 generate `adult' (a) ewe parameters, whereas female (f) and male (m) 16 month old parameters were called `hogget' (h) parameters (Table 1). The WAp&g were taken from Greeff (1997) and Greeff (pers. comm.). The `permissibility' of the resulting phenotypic and genetic variance-covariance matrices was tested (Hill and Thompson 1978; Foulley and Olivier 1986) and the necessary conditions were satisfied for both SAp&g and WAp&g. Table 1. SA (bold) and WA parameters : phenotypic standard deviations (CJP), heritabilities (along diagonal) and phenotypic (above) and genetic (below) correlations hCFWm CP hCFW m hCFWf aCFWf hFDm hFDf aFDf hCVm hCVf aCVf hSSm hSSf aSSf 15.80 15.53 0.57 :;iA 0.73 0.38 0.25 0.21 0.21 -0.08 0.16 0.00 -o.fB 0.44 0.21 0.13 8.14 hCFWf 15.00 aCFWf 16.00 hFDm 1.50 1.75 5 '' O.62 0.50 0.96 0.93 -0.24 -0.04 -0.16 -09 0.50 0.24 o.so 0.43 hFDf 1% aFDf 1.76 hCVm 2.40 2.57 hCVf 2.40 aCVf 2.40 hSSm 11.5 8.91 (124 '06 0.33 0.16 -0.17 -0.01 0.60 0.51 0.97 0.83 -0.42 -0.12 -0.40 -0.55 -0.11 -0.21 -0.35 -0.49 0.71 0.84 -0.48 -0.56 -0.52 0.65 0.66 A-J.45 -0.50 -0.52 -0.31 -0.33 0.45 0.41 0.59 0.65 -0.24 -0.49 0.27 0.23 0.20 0.20 hSSf 9.10 aSSf 10.5 0.42 0.80 0.07 0.30 0 59 0:45 0.13 0.10 0.26 0.15 0.10 0.04 O.10 o,ds 0.16 037 0.10 0 24 0:23 ' -0.16 -0.09 -0 01 -O:e3 -0 06 0.;3 0 10 0:11 0 07 ok ot37 -0.03 -0.08 -9.17 0.1) 0109 0.16 0.72 0.92 -0.15 -0.2Q -0.19 0.27 0.43 0.53 0.75 0.70 -0.04 -0.08 -0.03 0.30 0.33 0.45 0.42 0.40 0.45 0.35 * The WAp&g do ndt distinguish between the sheep classes, however for convenience the parametersare placed in the hogget sections of this Table. A base breeding objective (BASE) was defined which included CFW and FD. The breeding objective was then Iexpanded to include CV and SS. The SAP&g distinguish between the hogget male, the hogget female and the adult female expressions of these traits. The expanded genetic model was used because the genetic correlation between trait expressions in the two sexes or at different ages was not equal to one (Table 1). By contrast, the WAp&g (and reports on genetic change based on these parameters) treat the expression of these traits in the different sexes and ages as if they were the s me trait. The economic values were calculated (Ponzoni 1988) for two different micron premiums (3'and 12 o/e), and for three different price differentials of staple strength ($0.03, 0.06 and 0.12 per Newton per kilotex per kg of clean wool), assuming the price of 1 kg of clean wool was $4.50. Genetic change was calculated for a standard selection index which included hCFWm and hFDm, and then for indices which included hCVm and hSSrn. The genetic change was calculated for a period of 10 years, assuming the ratio of average selection intensity to generation intervals (in males and females) was 0.4. 47 Proc. Assoc. Advmt. Anim. Breed. Genet. Vol13 RESULTS AND DISCUSSION Table 2 shows that with the WA approach the inclusion of CV in the breedii objective and in the selection index was enough to stop the deterioration of SS at all micron premituns. That was not the case when SAp$g and a more elaborate genetic model were used. At a mkxm premium of 12 % there was a reduction in SS with the SAP&g even when SS was in the breedilpg objwctive (with low and medium economic values) but not as a selection criterion. When SS was a selection criterion the deolir#z still occurred at low a&medium economic value&k SS, but was aom&uha&at&nuated: Table 2, Predicted genetic changes from the use of SAP&g (bold) and WAphg B.Obj. 3 BASE BASE.+ CV BASE + CV +ssL BASE +SSM BASE +ssn BASh + SSL +CV +CV +CV BASE BASE -XV BASE cv BASE CV BASE cv 20.10 ?ZOIA 20`44 22.12 21.52 20.74 20.36 17.97 15.50 13.19 2!ho 21.46 19.06 18.64 13.53 14.48 9.80 4.64 10.23 340 11.09 3.75 11.77 3.94 12.29 3.99 10:82 5.29 11.21 6.07 10.99 6.74 20.87 19.76 19.97 18.15 12.75 19.70 17.36 11.59 -2.02 -0.64 -1.93 -0.63 -1.00 -0.23 -8.03 0.09 1.30 0.45 -1.04 -0.05 -0.11 0.26 1.08 0.56 -3.60 -3.01 -3.44 73.02 -3.19 -2.74 -2.85 -2.41 -1.91 -1.72 -3121 -2.58 -2.86 -2.13 -1.90 -1.33 -2.24 -2.19 -1.26 -0.26 1.15 -1.25 -0.25 1.06 0.75 0.83 O.Q4 0.95 -1.55 -0J35 -2.98 -2.12 -4.28 -3.39 -l&2 -1.22 -2.94 -2.32 -4.15 -3.30 1.07 m7 -8.m -Q 80. ,. -1.48 -1.71 -2.30, -2.48 -3.86 -3.53 -1;53 -1.91 -X36 -2.67 -3.82 -3.57 0.79 WI -1.25 -2M -3.79 -l&l -252 -3.66 Set. I+: CFW ('/) Hognet Adult FDtw Homet , Adult CVi%),, :i Adult Hogget 7.` &S (Nktex) 'H&get Adult 1. fb2 46 ~3.61 . .I. t9 dw3 5.46 3.35 ,&19 7.43 10.81 Q7!3 `4.36 4.66 12.24 8.IIfJ i4.38 -8.49 `-3.57 -2,29 1.88 5.61 9.82 2.74 7.03 11.35 + + + BASE +cv +ss BASE +cV +SS BASE +cv +ss BASE +CV + SSM BASE +CV + SSB 12 % micron premium BASE BASE BASE + CV BASE + CV +ssL BASE + CV +SSM BASE + CV +SSH BASE + C'? +ssL BASE + CV +SSM BASE + CV +SSH BASE +CV BASE cv BASE CV BASE CV BASE cv+ss BASE cV+ss BASE cv+ss 8.34 7.86 7.33 7-45 7.14 7.25 7.26 6.63 -4.02 -3.98 -3.76 j.44 -2.49 -3.94 -3.38 -2.35 1.39 -8:87 -7.40 4.33 -2.63 .ms : &@8~1. 0.62 -tzt -3.h -0.72 3.38 s -1.97 O;& 5.14 + + + + + + `-OsT5 -1.4'! -2.89 -o.sp -1.53 12.88 ihs `b.06 S.&J `4.46 j5.34 -2.46 8.:23 I:* 11.77 * The genetic changes derived from the WAp&g are placed in the hogget sections of this Table. 48 Proc. Assoc. Advmt. Anim. Breed. Genet. Voll3 Overall the results show a striking contrast between the predicted genetic ohanges in SS SAp&g and the WAp&g. This may be attributed to the differences in the genetic models as well as to the actual phenotypic and genetic parameter values used. Both differences genetic model and the parameter values) contribute towards reduced expectations about the of improving SS by genetic means using the South Australian approach. using the assumed, (i.e. the prospects Note that using the SAP&g approach genetic gains were greater (or losses smaller) for aSSf than for hSSf. This was due to a combination of factors, namely, greater economic value for aSSf than for hSSf, and stronger correlations of hFDCVm and hSSm with aSSf than with hSSf. When these values were `smoothed' (i.e. hFDCVm with hSSf and aSSf set equal to -0.45, and hSSm with hSSf and aSSf set equal to 0.6) the differences in genetic change between hSSf and aSSf were smaller, but still in favour of the latter trait. However, the overall conclusions drawn from the study remained unchanged The results based on the SAp&g suggest that although there is scope for genetic improvement of SS in Australian Merino sheep, gains are likely to be smaller and harder to achieve than earlier suggested by Western Australian studies based on WAp&g and on an over-simplified breeding objective. We conclude that the elucidation of an appropriate genetic model and the choice of the most appropriate phenotypic and genetic parameters are critical if realistic predictions of genetic change in SS are to be made. ACKNOWLEDCMENTS Financial assistance for the work was provided by Australian woolgrowers through the Woolmark Company, and J.A. Hill is the holder of a scholarship funded by SARDI. We thank Mr R.J. Grimson, Mrs KS. Jaensch and Mr D.H. Smith for their assistance. REFERENCES Foulley, J.L. and Qlivier, L. (1986) J. Anim. Breed. Genei. 103: 81 Gilmour, A.R., Cullis, B.R., Welham, S., and Thompson, R. (1998) Ag. Biom; Blrll.' No. 3. (Orange, NSW, Australia) Greeff, J.C., Lewer, R.P., Ponzoni, R.W. and Purvis, I.W. (1995) Proc Aust. Assoc. Anim. Breed. Genet. 11: 595 Greeff, J.C., Ritchie, A.J.M. and Lewis, R.M. (1997) Proc. Aust. Assoc. Anim. Breed..Genet. 12: 714 Greeff, J.C. (1997) In 'Access to the Experts; Module 1'. (CRC for Premium Quality Wool) Hill, W.G. and Thompson, R. (1978) Biometrica 34: 42' Lewer, R: and Li, Y. (1994) Wool Tech. Sheep Breed. 42: 103 Ponzoni, R.W. (1988) Wool. Tech. Sheep Breed, 36: 70 Ponzoni, R.W., Grimson, R.J., Jaensch, K.S., Smith, D.H., Gifford, D.R., Ancell, .P.M.C., Walkley, J.R.W. and Hynd, P.I. (1995) Proc. Astst. Assoc. Anim. Breed. Genet., 11:,303 49