Across population genetic parameters for wool, growth, and reproduction traits in Australian Merino sheep. 1. Data structure and non-genetic effects

CSIRO PUBLISHING www.publish.csiro.au/journals/ajar Australian Journal of Agricultural Research, 2007, 58, 169 175 Across population genetic parameters for wool, growth, and reproduction traits in Australian Merino sheep. 1. Data structure and non-genetic effects E. Safari A,G,N.M.Fogarty A, A. R. Gilmour A, K. D. Atkins A, S. I. Mortimer B,A.A.Swan C, F. D. Brien D, J. C. Greeff E, and J. H. J. van der Werf F A The Australian Sheep Industry Cooperative Research Centre, NSW Department of Primary Industries, Orange Agricultural Institute, Orange, NSW 2800, Australia. B NSW Department of Primary Industries, Agricultural Research Centre, Trangie, NSW 2823, Australia. C CSIRO Livestock Industries, Armidale, NSW 2350, Australia. D South Australian Research and Development Institute, Roseworthy, SA 7371, Australia. E Department of Agriculture and Food, Western Australia, Great Southern Agricultural Research Institute, Katanning, WA 6317, Australia. F School of Rural Science and Agriculture, University of New England, Armidale, NSW 2351, Australia. G Corresponding author. Email: alex.safari@dpi.nsw.gov.au Abstract. Accurate estimates of adjustment factors for systematic environmental effects are required for genetic evaluation systems. This study combined data from 7 research resource flocks across Australia to estimate genetic parameters and investigate the significance of various environmental factors for production traits in Australian Merino sheep. The flocks were maintained for several generations and represented contemporary Australian Merino fine, medium, and broad wool bloodlines over the past 30 years. Over 110 000 records were available for analysis for each of the major wool traits, with over 2700 sires and 25 000 dams. Univariate linear mixed animal models were used to analyse 6 wool, 4 growth, and 4 reproduction traits. This first paper outlines the data structure and the non-genetic effects of age of the animal, age of dam, birth-rearing type, sex, flock, bloodline, and year, which were significant with few exceptions for all production traits. Age of dam was not significant for reproduction traits and fleece yield. Generally, wool, growth, and reproduction traits need to be adjusted for age, birth-rearing type, and age of dam before the estimation of breeding values for pragmatic and operational reasons. Adjustment for animal age in wool traits needs to be applied for clean fleece weight (CFW), greasy fleece weight (GFW), and fibre diameter (FD) with inclusion of 2 age groups (2 years old and >2 years old), but for reproduction traits, inclusion of all age groups is more appropriate. For GFW, CFW, and hogget weight (HWT), adjustment for only 2 dam age groups of maiden and mature ewes seems sufficient, whereas for birth (BWT), weaning (WWT), and yearling (YWT) weights, adjustments need to be applied for all dam age groups. Adjustment for birth-rearing type (single-single, multiple-single, multiple-multiple) is appropriate for wool, growth, and reproduction traits. The implications of adjustment for non-genetic effects are discussed. Additional keywords: adjustment factors, dam age, birth-rearing type, age. Introduction The recent compilation of genetic parameters in sheep (Safari and Fogarty 2003) and subsequent summary and review (Safari et al. 2005) showed that there were more than 20 independent estimates of heritability for the major wool and growth traits, with small sampling variances around the means. There were considerably fewer estimates of heritability for reproduction, carcass, and disease traits. Safari et al. (2005) also showed that mean genetic correlations among the various production traits were typically based on few estimates and were associated with wide confidence intervals, especially between trait groups. Development of effective genetic evaluation and improvement programs requires knowledge of the genetic parameters and environmental effects that require adjustment for the economically important production traits. The parameters need to be estimated from relevant populations as parameters and fixed effects may vary among breeds and different populations. It is also important in parameter estimation that other variance components such as maternal and permanent environmental effects be included in models and assessed for their importance as they may bias estimates of direct genetic effects (Clément et al. 2001; Maniatis and Pollott 2003). In a further study, Safari et al. (2006) reported the sensitivity of selection response to changes in genetic correlations between some production traits. These results stressed the urgent need to obtain precise estimates of genetic parameters, CSIRO 2007 10.1071/AR06161 0004-9409/07/020169

170 Australian Journal of Agricultural Research E. Safari et al. especially genetic correlations between different production trait groups. Very large datasets, with appropriate structure and extensive pedigrees, are required to obtain accurate estimates of genetic correlations and other variance components. The study described in this series of papers has, for the first time, combined data from 7 large research resource flocks that represent the major strains and bloodlines of Merino sheep used in the Australian industry. Genetic analyses have been undertaken on the amalgamated dataset to provide accurate estimates of parameters for a range of wool, growth, and reproduction traits. This paper is the first in a series and it describes the data, their structure, and the effect of non-genetic factors on wool, growth, and reproduction traits derived from analyses of these combined data. Materials and methods Data The data were obtained from 7 Merino research resource flocks established by 4 different organisations over the past 30 years and are representative of the major strains of Merino sheep in the Australian population. The flocks include the Trangie D flock, Trangie C flock, Trangie QPLU$ flock (QPLU$), CSIRO fine wool flock, South Australian Base flock (SABASE), South Australian Selection Demonstration flock (SASDF), and Western Australian (WA) flock. The flocks were established for varying purposes including assessment of variation between and within strains and bloodlines, parameter estimation, evaluation of heterosis, and assessment of response to selection. In all flocks, a comprehensive range of production traits was recorded with full pedigrees of animals for several generations. The primary objectives and a brief description of the flocks together with references to further details are summarised in Table 1. The WA flock was restructured in 1998 into different selection lines for wool, staple strength, and meat traits. Summary statistics for the various flocks are presented in Table 2 for the wool traits: clean (CFW) and greasy (GFW) fleece weight, mean fibre diameter (FD), clean yield (YLD), coefficient of variation of fibre diameter (CVFD), and standard deviation of fibre diameter (SDFD). Similar statistics for the growth traits: liveweight at birth (BWT), weaning (WWT), yearling (YWT, 10 13 months), and hogget (HWT, 14 17 months) ages are shown in Table 3 and for the ewe reproduction traits: fertility (FER, ewes lambing per ewe joined), litter size (LS, lambs born per ewe lambing), lambs born per ewe joined (LB/EJ), and lambs weaned per ewe joined (LW/EJ) in Table 4. The summary statistics and the numbers of records, sires, and dams for the various production traits in the combined dataset are shown in Table 5. There were similar numbers of males and females for BWT and more females for the other growth traits. Wethers comprised 4% of the records for WWT and 1% for HWT and YWT. For the wool traits, 78% of records were from ewes with approximately 4 records per ewe, whereas rams generally had only 1 or 2 records. Ewes were generally joined to lamb for the first time at about 2 years of age and had approximately 4 joining and lambing records for reproduction traits. Table 1. Description of the Merino research resource flocks Flock Objective and description Years Location Organisation Reference Trangie D Variation within and between bloodlines: 2 fine, 2 medium non-peppin, 9 medium Peppin, 1 broad, 1 Fertility bloodlines Trangie C 8 8 diallel cross of Trangie D flock bloodlines: 2 fine, 2 medium non-peppin, 3 medium Peppin, 1 broad bloodline QPLU$ 9 selection lines for a range of indices combining high fleece weight and low fibre diameter among fine, medium, and broad wool strains (control, industry, and 3, 8, 15% micron premium objectives) CSIRO Variation within and between industry bloodlines: 9 fine and 2 medium wool SABASE Estimation of genetic parameters for wool and skin traits in 4 medium and broad wool bloodlines SASDF Selection using alternative breeding strategies in 4 medium and broad wool bloodlines WA Genetic variation in wool traits in 4 medium and broad wool bloodlines, with selection lines for wool and meat established in 1998 1975 89 Trangie, NSW NSW Department of Primary Industries 1984 95 Trangie, NSW NSW Department of Primary Industries 1993 2003 Trangie, NSW NSW Department of Primary Industries Mortimer and Atkins (1989) Mortimer et al. (1994) Taylor and Atkins (1997) 1990 2001 Armidale, NSW CSIRO Livestock Industries Swan et al. (2000) 1989 97 Turretfield, SA SA Research and Development Institute 1997 2004 Turretfield, SA SA Research and Development Institute 1984 2004 Katanning, WA Department of Agriculture and Food Western Australia Ponzoni et al. (1995) Ponzoni et al. (1999) Lewer et al. (1992)

Genetic parameters in Merino sheep. 1 Australian Journal of Agricultural Research 171 Table 2. Number of records (N), mean (s.d.), and coefficient of variation (CV) for clean fleece weight (CFW), greasy fleece weight (GFW), fibre diameter (FD), clean yield (YLD), coefficient of variation of fibre diameter (CVFD), and standard deviation of fibre diameter (SDFD) recorded in 7 Merino research resource flocks Trait Trangie D Trangie C QPLU$ CSIRO SABASE SASDF WA CFW N 17 692 16 015 27 798 13 915 12 876 5801 21 147 Mean (s.d.) kg 3.6 (0.81) 3.6 (0.79) 4.3 (1.15) 2.7 (0.73) 4.5 (1.09) 5.0 (0.91) 3.5 (0.80) CV % 22 22 28 26 31 18 22.5 GFW N 18 058 16 116 29 022 14 002 12 899 5806 21 895 Mean (s.d.) kg 5.2 (1.03) 5.2 (1.04) 6.1 (1.53) 3.5 (0.87) 6.2 (1.44) 6.9 (1.13) 4.95 (1.10) CV % 20 20 25 25 23 19 22.2 FD N 17 664 15 993 27 829 13 374 12 927 5817 22 421 Mean (s.d.) µm 21.8 (2.06) 21.1 (2.05) 21.6 (2.40) 18.4 (1.60) 23.9 (2.40) 20.6 (1.98) 21.2 (2.25) CV % 9.4 9.7 11.1 8.7 10.2 9.6 10.6 YLD N 17 690 16 014 27 830 13 916 12 889 5817 22 370 Mean (s.d.) % 68.9 (6.40) 69.9 (5.72) 72.2 (5.85) 76.9 (4.40) 72.5 (5.57) 72.8 (5.46) 70.8 (5.15) CV % 9.3 8.2 8.1 5.7 7.7 7.5 7.3 CVFD N 2914 27 829 13 374 12 927 858 18 701 Mean (s.d.) % 20.2 (3.14) 20.9 (2.90) 16.8 (2.21) 22.8 (2.84) 20.0 (2.52) 22.3 (3.23) CV % 15.5 13.9 13.2 12.5 12.6 14.5 SDFD N 2914 27 825 12 927 5817 6452 Mean (s.d.) µm 4.32 (0.80) 4.51 (077) 5.42 (0.74) 4.53 (0.68) 4.70 (0.77) CV % 18.5 17.1 13.6 15.0 16.4 Table 3. Number of records (N), mean (s.d.), and coefficient of variation (CV) for birth weight (BWT), weaning weight (WWT), yearling weight (YWT), and hogget weight (HWT) recorded in 7 Australian Merino research resource flocks Trait Trangie D Trangie C QPLU$ CSIRO SABASE SASDF WA BWT N 11 786 9704 13 959 6590 6734 4968 19 399 Mean (s.d.) kg 3.62 (0.80) 4.06 (0.73) 4.18 (0.87) 4.22 (0.72) 4.55 (0.92) 4.88 (0.91) 4.71 (0.89) CV % 22.0 18.0 20.8 17.1 20.2 18.6 18.9 WWT N 9395 9714 13 815 6573 9536 4962 18 343 Mean (s.d.) kg 19.1 (4.4) 21.8 (4.4) 21.3 (4.8) 16.7 (3.3) 21.9 (5.7) 26.2 (5.5) 24.8 (5.7) CV % 23.0 20.2 22.5 19.8 26.0 21.0 23.0 YWT N 8620 9678 5016 4947 Mean (s.d.) kg 30.6 (6.8) 33.3 (6.7) 25.1 (4.2) 41.5 (7.7) CV % 22.2 20.1 16.7 18.5 HWT N 3766 9654 12 978 4276 5800 16 793 Mean (s.d.) kg 38.3 (7.4) 46.4 (8.5) 53.7 (10.4) 34.1 (5.1) 51.4 (7.3) 50.0 (9.3) CV % 19.3 18.3 19.4 15.0 14.2 18.7 Table 4. Number of records (N), mean (s.d.), and coefficient of variation (CV) for fertility (FER), litter size (LS), lambs born per ewe joined (LB/EJ), and lambs weaned per ewe joined (LW/EJ) recorded in 7 Australian Merino research resource flocks Trait Trangie D Trangie C QPLU$ CSIRO SABASE SASDF WA FER N 14 418 6534 14 797 9282 10 636 783 12 937 Mean (s.d.) 0.73 (0.44) 0.84 (0.36) 0.77 (0.42) 0.80 (0.40) 0.81 (0.39) 0.86 (0.34) 0.87 (0.34) CV % 60.3 42.9 54.5 50.0 48.1 39.5 39.1 LS N 10 505 5512 11 270 7452 8636 677 11 209 Mean (s.d.) 1.40 (0.52) 1.42 (0.52) 1.49 (0.54) 1.07 (0.26) 1.39 (0.51) 1.18 (0.39) 1.30 (0.47) CV % 37.1 36.6 36.2 24.3 36.7 33.0 36.1 LB/EJ N 14 418 6534 14 797 9282 10 636 783 12 937 Mean (s.d.) 1.02 (0.76) 1.19 (0.70) 1.13 (0.79) 0.86 (0.49) 1.12 (0.71) 1.02 (0.55) 1.13 (0.62) CV % 76.0 58.3 71.8 57.0 64.5 55.0 56.4 LW/EJ N 14 418 6534 14 797 9282 10 636 783 12 937 Mean (s.d.) 0.75 (0.70) 0.92 (0.71) 0.82 (0.75) 0.71 (0.52) 0.90 (0.68) 0.79 (0.54) 0.93 (0.62) CV % 93.3 77.2 91.5 73.2 75.6 71.0 66.7

172 Australian Journal of Agricultural Research E. Safari et al. Table 5. Mean (s.d.) and numbers of records, sires, and dams for traits in the combined dataset Trait A Mean (s.d.) Records Sires Dams CFW 3.83 (1.11) 115 244 2707 25 160 GFW 5.3 (1.50) 117 798 2707 25 168 FD 21.3 (2.55) 116 025 2707 25 168 YLD 71.7 (6.02) 116 526 2707 25 165 CVFD 20.8 (3.45) 76 603 1874 18 042 SDFD 4.7 (0.85) 55 935 1607 14 362 BWT 4.30 (0.92) 73 140 2713 27 977 WWT 21.9 (5.52) 72 338 2791 28 072 YWT 32.5 (6.55) 28 261 1296 11 150 HWT 48.2 (9.54) 52 475 2378 22 422 FER 0.80 (0.40) 69 388 2312 14 379 LS 1.35 (0.50) 55 260 2300 13 608 LB/EJ 1.08 (0.70) 69 388 2312 14 379 LW/EJ 0.83 (0.67) 69 388 2312 14 379 A See Tables 2, 3, and 4 for explanation of trait codes. Statistical analysis A linear mixed animal model was used for the analysis. Direct and maternal genetic, animal, and maternal permanent environmental, litter, and residual effects were included as random effects in the model. The fixed effects included in the model were dam age (6 levels: 2 to 7), birth-rearing type (3 levels: single-single, multiple-single, multiple-multiple), sex (3 levels: male, female, wether), age (7 levels, 1 to 7), flock (7 levels), bloodline (67 levels), and flock/year of record as management group. Weaning age was included as a covariate for weaning weight, yearling weight, and hogget weight. The interactions of management group with the other effects were also included. Those interactions that were not significant (P > 0.05) were removed from the final model. The analyses were carried out using ASReml (Gilmour et al. 2002). Results Flocks The flock differences in production levels for the various traits reflect both the genetic effects of different strains and merit of individuals and the environmental effects of the very different regions in which the flocks were located, the years of data collection, and management effects. The differences between the strains represented in the flocks are illustrated by the high wool production in the SABASE and low wool production in the CSIRO flock and the contrasting levels for FD (Table 2). The coefficients of variation for wool traits were generally similar across the flocks, with a slightly higher value for FD in QPLU$ reflecting the inclusion of fine, medium, and broad wool strains. Similarly, for liveweight traits the CSIRO flock with mainly fine wool bloodlines generally had lower means and coefficients of variation across ages in comparison with other flocks. In addition to the influence of strain and bloodline, the differences between liveweights among the flocks (Table 3) are also due to differences in ages at recording, management, and environmental conditions. There was a range in FER (0.73 0.87) among the flocks, which also resulted in differences for the composite reproduction traits of LB/EJ (0.86 1.19) and LW/EJ (0.71 0.93). There was a large range in LS (1.07 1.49), which could reflect both strain and environmental effects (including management and age of measurement). The coefficients of variation were generally similar across the flocks for each of the reproduction traits (Table 4). The summary statistics and the numbers of records, sires, and dams for the various production traits in the combined data set are shown in Table 5. Wool traits Predicted means and their standard errors for age, birth-rearing type, and dam age are presented in Table 6. All the fixed effects were highly significant (P < 0.001) with the exception of dam age for YLD. The interaction of sex with management group was significant (P < 0.001). Both animal age and dam age effects were curvilinear (quadratic) except for FD where a linear relationship was observed. Predicted means of CFW, GFW, and YLD had a similar age pattern, with a maximum at 4 years for CFW and GFW and at 3 years for YLD then a gradual decline at older ages. On the other hand, FD increased up to 6 years of age. CVFD and SDFD declined up to 5 years of age and then increased during the next 2 years. The means for birth-rearing type showed a declining trend in CFW and GFW from single to multiple. In contrast, FD, CVFD, and SDFD increased from single to multiple. Multiple-born and single-reared animals had lower mean YLD than single-single or multiple-multiple animals. There was an increase in CFW and GFW with dam age up to 5 years, with a decline thereafter, in contrast to CVFD and SDFD where there was an increasing trend observed across dam age groups. Growth traits All the fixed effects were highly significant (P < 0.001), with a quadratic effect for dam age (Table 7). The interaction of sex with management group was significant (P < 0.001) for all weights. The effect of dam age was similar for all growth traits, with an increasing trend for dam age up to 6 years for BWT and 5 years for other weights. Single-born and -reared lambs were heavier than multiple-multiple at all ages, with multiple-single lambs being intermediate. Males were heavier than females at birth. Sex effects are not presented for WWT, YWT, or HWT because of the confounding of sex and management effects. Reproduction traits All the fixed effects except dam age were highly significant (P < 0.001) for the reproduction traits (Table 8). The interaction of sex with management group was significant (P < 0.001). The age effect was quadratic, with FER, LB/EJ, and LW/EJ increasing up to 5 years and then declining, whereas LS reached a plateau at 6 years. There was a large increase in performance for ewes from 2 (first joining) to 3 years of age for all reproduction traits. The performance of ewes that had been born and reared as multiples was significantly (P < 0.01) higher than of those born and reared as singles, with multiple-single intermediate for all reproduction traits. Discussion Identification of superior animals and subsequent selection decisions should be based on genetic merit rather than on differences due to environmental effects. Hence, performance

Genetic parameters in Merino sheep. 1 Australian Journal of Agricultural Research 173 Table 6. Predicted means (± s.e.) for fixed effects for clean fleece weight (CFW), greasy fleece weight (GFW), fibre diameter (FD), yield (YLD), coefficient of variation of fibre diameter (CVFD), and standard deviation of fibre diameter (SDFD) Within columns, means followed by the same letter are not significantly different at P = 0.05 CFW GFW FD YLD CVFD SDFD Age A *** *** *** *** *** *** 1 3.38 ± 0.01a 4.75 ± 0.05a 20.76 ± 0.03a 71.40 ± 0.07a 21.51 ± 0.05a 4.89 ± 0.02a 2 3.75 ± 0.01b 5.18 ± 0.05b 21.40 ± 0.03b 71.83 ± 0.07b 20.45 ± 0.05b 4.77 ± 0.01b 3 3.96 ± 0.01c 5.43 ± 0.05c 21.88 ± 0.03c 72.01 ± 0.07c 19.73 ± 0.05c 4.69 ± 0.01c 4 4.00 ± 0.01d 5.48 ± 0.05d 22.22 ± 0.03d 71.93 ± 0.08d 19.35 ± 0.05d 4.65 ± 0.02d 5 3.89 ± 0.01e 5.35 ± 0.05e 22.43 ± 0.03e 71.58 ± 0.08e 19.30 ± 0.06e 4.65 ± 0.02d 6 3.61 ± 0.01f 5.02 ± 0.05f 22.49 ± 0.04f 70.98 ± 0.09f 19.58 ± 0.07f 4.70 ± 0.02e 7 3.17 ± 0.02g 4.51 ± 0.05g 22.42 ± 0.04g 70.13 ± 0.12g 20.20 ± 0.09g 4.79 ± 0.02f Birth-rearing type B *** *** *** *** *** *** Single-single 4.13 ± 0.01a 5.66 ± 0.05a 22.12 ± 0.03a 72.03 ± 0.08a 19.22 ± 0.05a 4.61 ± 0.02a Multiple-single 3.95 ± 0.01b 5.42 ± 0.05b 22.23 ± 0.03b 71.75 ± 0.09b 19.36 ± 0.06b 4.66 ± 0.02b Multiple-multiple 3.92 ± 0.01c 5.37 ± 0.05c 22.32 ± 0.03c 71.99 ± 0.08c 19.46 ± 0.05c 4.69 ± 0.02c Dam age C *** *** ** n.s. *** *** 2 3.89 ± 0.01a 5.33 ± 0.05a 22.19 ± 0.03ab 71.86 ± 0.08a 19.32 ± 0.06a 4.63 ± 0.02a 3 3.96 ± 0.01b 5.42 ± 0.05b 22.17 ± 0.03a 71.87 ± 0.08a 19.33 ± 0.05a 4.64 ± 0.02a 4 4.00 ± 0.01c 5.48 ± 0.05c 22.22 ± 0.03b 71.95 ± 0.08a 19.35 ± 0.05a 4.65 ± 0.02b 5 4.02 ± 0.01d 5.51 ± 0.05d 22.18 ± 0.03a 71.94 ± 0.07a 19.38 ± 0.05a 4.66 ± 0.01c 6 4.01 ± 0.01d 5.49 ± 0.05e 22.17 ± 0.03a 72.00 ± 0.08a 19.43 ± 0.05c 4.68 ± 0.01d 7 3.98 ± 0.01e 5.45 ± 0.05e 22.11 ± 0.03c 71.94 ± 0.09a 19.50 ± 0.06d 4.69 ± 0.02d ***P < 0.001; **P < 0.01; n.s., not significant. A Adjusted to dam age of 4 years. B Adjusted to age 4 years and dam age 4 years. C Adjusted to age 4 years. Table 7. Predicted means (± s.e.) for fixed effects for birth weight (BWT), weaning weight (WWT), yearling weight (YWT), and hogget weight (HWT) Within columns, means followed by the same letter are not significantly different at P = 0.05 BWT WWT YWT HWT Birth-rearing type A *** *** *** *** Single-single 4.75 ± 0.01a 23.38 ± 0.05a 30.46 ± 0.11a 47.42 ± 0.10a Multiple-single 3.92 ± 0.01b 21.67 ± 0.06b 29.41 ± 0.12b 46.33 ± 0.12b Multiple-multiple 3.42 ± 0.03c 19.45 ± 0.05c 28.16 ± 0.11c 45.54 ± 0.10c Sex A *** *** *** *** Male 4.15 ± 0.01a Female 3.91 ± 0.01b Dam age *** *** *** *** 2 3.69 ± 0.01a 20.36 ± 0.06a 28.37 ± 0.11a 45.37 ± 0.11a 3 3.89 ± 0.01b 21.12 ± 0.05b 28.99 ± 0.10b 46.03 ± 0.11b 4 4.03 ± 0.01c 21.54 ± 0.05c 29.35 ± 0.10c 46.43 ± 0.10c 5 4.11 ± 0.01d 21.62 ± 0.05d 29.44 ± 0.10d 46.58 ± 0.10d 6 4.12 ± 0.01d 21.36 ± 0.05e 29.27 ± 0.10e 46.46 ± 0.10e 7 4.07 ± 0.02e 20.76 ± 0.07f 28.84 ± 0.13f 46.09 ± 0.13f ***P < 0.001. A Adjusted to dam age 4. records of animals need to be adjusted for the non-genetic sources of variation either before or during the process of estimation of breeding values. As pointed out by Notter et al. (2005), adjustment factors external to the data are preferred for pragmatic and operational reasons and they are applied in many genetic evaluation systems, including those used in Australia (Brown et al. 2000). Selection was applied in some of the resource flocks and data were collected over a period of 29 years. Hence the predicted means for environmental factors were obtained using models that included all significant random effects to avoid introduction of bias due to possible genetic and environmental trends (Lush and Shrode 1950). LAMBPLAN, which has now been incorporated into Sheep Genetics Australia, the national sheep evaluation program, adjusts phenotypic records for dam age, animal age, and birth rearing type (Brown et al. 2000). Given the large number of available records within each subclass of environmental factors, accurate means were

174 Australian Journal of Agricultural Research E. Safari et al. Table 8. Predicted means (± s.e.) for fixed effects for fertility (FER), litter size (LS), lambs born per ewe joined (LB/EJ), and lambs weaned per ewe joined (LW/EJ) Within columns, means followed by the same letter are not significantly different at P = 0.05 FER LS LB/EJ LW/EJ Age *** *** *** *** 2 0.83 ± 0.03a 1.17 ± 0.03a 0.99 ± 0.05a 0.76 ± 0.08a 3 0.88 ± 0.03b 1.29 ± 0.03b 1.14 ± 0.05b 0.90 ± 0.08b 4 0.90 ± 0.03c 1.38 ± 0.03c 1.23 ± 0.05c 0.98 ± 0.08c 5 0.90 ± 0.03c 1.44 ± 0.03d 1.28 ± 0.05d 1.01 ± 0.08d 6 0.88 ± 0.03d 1.47 ± 0.03e 1.28 ± 0.05d 0.98 ± 0.08e 7 0.86 ± 0.03e 1.47 ± 0.03e 1.23 ± 0.06e 0.89 ± 0.08f Birth-rearing type A *** *** *** *** Single-single 0.89 ± 0.03a 1.37 ± 0.03a 1.22 ± 0.049a 0.96 ± 0.08a Multiple-single 0.90 ± 0.03ab 1.38 ± 0.03ab 1.24 ± 0.050ab 0.98 ± 0.08ab Multiple-multiple 0.91 ± 0.03b 1.39 ± 0.03b 1.25 ± 0.050b 1.00 ± 0.08b ***P < 0.001. A Adjusted to age 4. predicted for each subclass from our analysis. In our study, all these factors significantly affected the production traits of Merino sheep. The significant sex management group interaction was due to sexes being generally managed differently after weaning. The absence of interaction between management groups and other effects means that the trends are the same across a range of very different genotypes and environments. There was a significant quadratic effect of dam age for most wool traits, with the greatest increase being for the progeny of mature compared with 2-year-old ewes. These differences were of the order of 2.6% for CFW and GFW, which was similar to the differences reported by Lax and Brown (1967), Gregory and Ponzoni (1981), and Mortimer and Atkins (1989). The birthrearing type effect on GFW, CFW, and FD has proved to be significant in other studies (Lax and Brown 1967; Walkley et al. 1987; Mortimer and Atkins 1989; Lewer et al. 1992; Yazdi et al. 1998). In our study, multiple born and reared animals produced 5% less CFW than singles, with 0.2 µm higher FD. This was similar to the effects reported by Brown et al. (1966), Turner et al. (1968), and File (1981), although Mortimer and Atkins (1989) found that multiples had 8% less wool and 0.4 µm higher FD than singles when measured as hoggets. The curvilinear effect of age of dam on growth traits peaked at 5 years for WWT, YWT, and HWT and at 6 years for BWT, which is consistent with other studies (Lewis et al. 1989; Yazdi et al. 1998; Cloete et al. 2002; Notter et al. 2005). As for fleece weight, the major effect of dam age on growth traits was between dam age 2 and 3. However for BWT, WWT, and YWT, the differences between subsequent dam age groups were greater than 1% except between dam age 4 and 5, suggesting that these traits need to be adjusted for all dam age groups. In contrast, for HWT, adjustment needs to be applied for only 2 age groups, i.e. 2-year-old and adult dams (>2), because the only noticeable difference (>1%) is between 2- and 3-year-old dams. Although there were significant differences among birth-rearing types for all growth traits, the magnitude of differences generally declined from birth to hogget age, which is consistent with the findings of Yazdi et al. (1998). This is a reflection of the decline in maternal effect with increasing age, especially after weaning. Several studies have reported similar curvilinear effects of ewe age on reproduction traits in Merinos (Turner and Dolling 1965; Mullaney and Brown 1970; Gregory et al. 1977) and other breeds (Dickerson and Glimp 1975; Hohenboken et al. 1976; Notter 2000). In the Merino, Turner and Dolling (1965) reported the peak for LB/EJ and LW/EJ at 6 and 7 years of age, respectively, and Mullaney and Brown (1970), using the same data source with more records, reported the peak age for both traits at 6 years. This is similar to our study where the maximum for LB/EJ and LW/EJ was at 5 6 and 5 years, respectively. Although the biggest differences between age groups of ewes was observed between 2 and 3 years, there were age differences among mature ewes, suggesting that when enough records are available for different age groups, age-specific adjustment factors are preferable, as was also concluded by Notter (2000). Implications Major production traits were significantly affected by all the environmental factors studied in this investigation. Therefore, genetic evaluation and improvement programs need to include these factors in the models used for the estimation of breeding values. Wool, growth, and reproduction traits need to be adjusted for age, birth-rearing type, and dam age before the estimation of breeding values for pragmatic and operational reasons. For the wool traits CFW, GFW, and FD, adjustment for age using 2 age groups (2 years old and >2 years old) seems sufficient, but for reproduction traits the inclusion of all age groups seems more appropriate. For CFW, GFW, and HWT, adjustment for 2 dam age groups (2-year-old and mature ewes) seems sufficient, whereas for BWT, WWT, and YWT, adjustment needs to be applied for all dam age groups. For wool and growth traits and LW/EJ the inclusion of all birth-rearing types seems appropriate for the adjustment, whereas for the other reproduction traits, adjustment for multiple and single seems sufficient. Acknowledgments Funding for this study was provided by the Commonwealth Government through the Australian Sheep Industry Cooperative Research Centre. We also gratefully thank the many other scientists and technical and support

Genetic parameters in Merino sheep. 1 Australian Journal of Agricultural Research 175 staff who have contributed to the management of the flocks and collected the data over many years, from the Agricultural Research Centre, Trangie, and NSW Department of Primary Industries; CSIRO Livestock Industries, Armidale; Turretfield Research Centre and the South Australian Research and Development Institute; the Great Southern Agricultural Research Institute, Katanning, and the Department of Agriculture and Food Western Australia. Contributions of sheep breeders and industry funding bodies such as Australian Wool Innovation and Meat and Livestock Australia and their predecessors over many years to the various flocks are also gratefully acknowledged. References Brown DJ, Tier B, Reverter A, Banks R, Graser HU (2000) OVIS: a multiple trait breeding value estimation program for genetic evaluation of sheep. International Journal of Sheep and Wool Science 48, 285 297. Brown GH, Turner HN, Young SS, Dolling CHS (1966) Vital statistics for an experimental flock of Merino sheep. 3. Factors affecting wool and body characteristics, including the effect of age of ewe and its possible interaction with method of selection. Australian Journal of Agricultural Research 17, 557 581. doi: 10.1071/AR9660557 Clément V, Bibe B, Verrier E, Elsen JM, Manfredi E, Bouix J, Hanocq E (2001) Simulation analysis to test the influence of model adequacy and data structure on the estimation of genetic parameters for traits with direct and maternal effects. Genetic Selection Evolution 33, 369 395. doi: 10.1051/gse:2001123 Cloete SWP, Scholtz AJ, Gilmour AR, Olivier JJ (2002) Genetic and environmental effects on lambing and neonatal behaviour of Dormer and SA Mutton Merino lambs. Livestock Production Science 78, 183 193. doi: 10.1016/S0301-6226(02)00117-3 Dickerson GE, Glimp HA (1975) Breed and effects of lamb production of ewes. Journal of Animal Science 40, 397 408. File GC (1981) Highly fertile Merinos and their nutritional management through pregnancy and lactation. International Journal of Sheep and Wool Science 29, 7 11. Gilmour AR, Gogel BJ, Cullis BR, Welham SJ, Thompson R (2002) ASReml user guide release 1.0. (VSN International Ltd: Hemel Hempstead, UK) Gregory IP, Ponzoni RW (1981) Genetic studies of South Australian Merino sheep. II. Environmental effects on wool and body traits at 15 16 month of age. Australian Journal of Agricultural Research 32, 657 667. doi: 10.1071/AR9810657 Gregory IP, Roberts EM, James JW (1977) Genetic improvement of meat sheep. 4. Effects of age of dam on productivity of Dorset and Border Leicester sheep. Australian Journal of Experimental Agriculture and Animal Husbandry 17, 735 740. doi: 10.1071/EA9770735 Hohenboken W, Corum K, Bogart R (1976) Genetic, environmental and interaction effects in sheep. I. Reproduction and lamb production in ewe. Journal of Animal Science 42, 299 306. Lax J, Brown GH (1967) The effects of inbreeding, maternal handicap and range in age on 10 fleece and body characteristics in Merino rams and ewes. Australian Journal of Agricultural Research 18, 689 706. doi: 10.1071/AR9670689 Lewer RP, Woolaston RR, Howe RR (1992) Studies on Western Australian Merino sheep. I. Stud, strain and environmental effects on hogget performance. Australian Journal of Agricultural Research 43, 1381 1397. doi: 10.1071/AR9921381 Lewis RM, Shelton M, Sanders JO, Notter DR, Pirie WR (1989) Adjustment factors for a 120-day weaning weight in Rambouillet range lambs. Journal of Animal Science 67, 1107 1115. Lush JL, Shrode R (1950) Changes in milk production with age and milking frequency. Journal of Dairy Science 33, 338 357. Maniatis N, Pollott GE (2003) The impact of data structure on genetic (co)variance components of early growth in sheep, estimates using animal model with maternal effects. Journal of Animal Science 81, 101 108. Mortimer SI, Atkins KD (1989) Genetic evaluation of production traits between and within flocks of Merino sheep. I. Hogget fleece weight, body weight and wool quality. Australian Journal of Agricultural Research 40, 433 443. doi: 10.1071/AR9890433 Mortimer SI, Atkins KD, Eissen J, Van Heelsum A, Burns AM, Isaac BR (1994) Effect of changing Merino ram source on average hogget production and wool quality levels and between-animal variability. International Journal of Sheep and Wool Science 42, 243 252. Mullaney PD, Brown GH (1970) Some components of reproductive performance of sheep in Victoria. Australian Journal of Agricultural Research 21, 945 950. doi: 10.1071/AR9700945 Notter DR (2000) Effects of ewe age and season of lambing on prolificacy in US Targhee, Suffolk, and Polypay sheep. Small Ruminant Research 38, 1 7. doi: 10.1016/S0921-4488(00)00144-9 Notter DR, Borg RC, Kuehn LA (2005) Adjustments of lamb birth and weaning weights for continuous effects of ewe age. Animal Science 80, 241 248. doi: 10.1079/ASC40760241 Ponzoni RW, Grimson RJ, Jaensch KS, Smith DH, Gifford DR, Ancell PMC, Walkley JRW, Hynd PI (1995) The Turretfield sheep breeding project: messages on phenotypic and genetic parameters for South Australian Merino sheep. Proceedings of the Australian Association of Animal Breeding and Genetics 11, 303 313. Ponzoni RW, Jaensch KS, Grimson RJ, Smith DH, Ewers AL, Ingham V (1999) South Australian Merino selection demonstration flocks: background and first hogget results. International Journal of Sheep and Wool Science 47, 83 94. Safari E, Fogarty NM (2003) Genetic parameters for sheep production traits: estimates from literature. Technical Bulletin 49, NSW Agriculture, Orange, Australia. www.sheep.crc.org.au/articles.php3?rc=145. Safari E, Fogarty NM, Gilmour AR (2005) A review of genetic parameter estimates for wool, growth, meat and reproduction traits in sheep. Livestock Production Science 92, 271 289. doi: 10.1016/ j.livprodsci.2004.09.003 Safari E, Fogarty NM, Gilmour AR (2006) Sensitivity of multi-trait index selection to changes in genetic correlations between production traits in sheep. Australian Journal of Experimental Agriculture 46, 283 290. doi: 10.1071/EA04232 Swan AA, Purvis IW, Piper LR, Lamb PR, Robinson GA (2000) The CSIRO fine wool project background objectives. In Fine wool 2000: Proceeding of a Symposium. 27 28 October 2000, Armidale. pp. 65 73. (CSIRO Livestock Industries, Armidale and The Woolmark Company, Melbourne) Taylor PJ, Atkins KD (1997) Genetically improving fleece weight and fibre diameter of the Australian Merino the Trangie QPLU$ Project. International Journal of Sheep and Wool Science 45, 92 107. Turner HN, Brown GH, Ford GH (1968) The influence of age structure on total productivity in breeding flocks of Merino sheep. II. Flocks with a fixed number of breeding ewes, producing their own replacements. Australian Journal of Agricultural Research 19, 443 475. doi: 10.1071/ AR9680443 Turner HN, Dolling CHS (1965) Vital statistics for an experimental flock of Merino sheep. II. The influence of age on reproductive performance. Australian Journal of Agricultural Research 16, 699 712. doi: 10.1071/AR9650699 Walkley JRW, Ponzoni RW, Dolling CHS (1987) Phenotypic and genetic parameters for lamb and hogget traits in a flock of South Australian Merino sheep. Australian Journal of Experimental Agriculture 27, 205 210. doi: 10.1071/EA9870205 Yazdi MH, Eftekhari-Shahroudi F, Hejazi M, Liljedahl LE (1998) Environmental effects on growth traits and fleece weights in Baluchi sheep. Journal of Animal Breeding and Genetics 115, 455 465. Manuscript received 12 May 2006, accepted 16 October 2006 http://www.publish.csiro.au/journals/ajar