Evaluating breeding objectives for sow productivity and production traits in Large White Pigs

Livestock Science 157 (2013) 9–19

Contents lists available at ScienceDirect

Livestock Science journal homepage: www.elsevier.com/locate/livsci

Evaluating breeding objectives for sow productivity and production traits in Large White Pigs B. Dube a,n, S.D. Mulugeta b, K. Dzama a a b

Department of Animal Sciences, Stellenbosch University, Private Bag X1, Matieland 7602, South Africa Animal Science Programme, North West University, Private Bag X2046, Mmabatho 2735, South Africa

a r t i c l e i n f o

abstract

Article history: Received 3 August 2012 Received in revised form 10 June 2013 Accepted 12 June 2013

the objective of the study was to develop and evaluate different breeding objectives for sow productivity and for production traits, using economic selection indices. Genetic parameters were generated using a repeated records model for sow productivity traits and a maternal effects model for production traits, in ASREML. Stochastic simulation models based on a hypothetical 100-sow model were performed for each line, i.e., a dam line and a terminal sire line, respectively, to derive economic values. The traits included in the study were number born alive (NBA), 21-day litter size (D21LS), 21-day litter weight (D21LWT), average daily gain (ADG), feed conversion ratio (FCR), age at slaughter (AGES), dressing percentage (DRESS), lean content (LEAN) and backfat thickness (BFAT). The economic values for LEAN and BFAT were derived using the partial differentiation of the profit function, while those for the other traits were derived using the partial budget approach. An economic value of a trait was the change in profit per unit genetic change in that trait. Breeding objectives were developed with a corresponding selection index, for improvement of that objective. Three combinations of breeding objectives and selection indices were developed for sow productivity traits, while there were 15 combinations for production traits. Responses to selection and economic return were computed for each combination to determine the most appropriate combination for the improvement of the breeding objective traits. The most appropriate index to improve sow productivity consisted of NBA and D21LWT. For production traits, the combination that consisted of a selection index with AGES, DRESS and BFAT, and the breeding objective ADG, DRESS, FCR and LEAN, was considered the most appropriate. Age at slaughter and BFAT were, respectively, included as indicator traits for ADG and LEAN. The recommended breeding objectives were sensitive to changes in economic values, indicating that economic values for breeding goal traits should be updated periodically to ensure proper weighting of traits, hence maximization of economic return. & 2013 Elsevier B.V. All rights reserved.

Keywords: Genetic parameters Economic return Genetic selection Response to selection

1. Introduction Commercial pig production in South Africa is mainly based on crossbreeding, with dam lines contributing to sow productivity and terminal sire lines contributing to

n

Corresponding author. Tel.: +27 21 808 3794. E-mail addresses: [email protected], [email protected] (B. Dube).

1871-1413/$ - see front matter & 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.livsci.2013.06.018

performance in production traits. The breeding programs thus aim to improve these traits by having separate genetic improvement programs for the respective lines. In the past, pig production in South Africa focused on input parameters, such as, feed intake and feed conversion efficiency, with little emphasis on output parameters, such as, carcass yield and quality (Visser, 2004). Profitability has been viewed as a logical unit of expression for the final evaluation of a pig enterprise (MacNeil et al., 1997). Thus,

10

B. Dube et al. / Livestock Science 157 (2013) 9–19

of late, consideration of sow productivity, input and output parameters has been central to the South African pig improvement programs, through multi-trait selection and selection indices to improve the profitability of pig enterprises (Visser, 2004). This has been driven by the need to shift from production to productivity and sustainability, which is a common feature of modern pig industries (Olesen et al., 2000). The use of selection index methodology to select for more than one trait by weighting EBVs with economic values has long been established in the livestock industry (Banga, 2009; Groen, 1990; Wolfova et al., 2007). The original selection index developed by Hazel (1943) generated phenotypic weights, which have been discouraged in genetic improvement programs because of the limitations outlined by Bourdon (1998). Thus, an index that utilizes EBVs weighted by economic values is ideal for modern genetic selection programs (Bourdon, 1998; Wolfova et al., 2007). Use of economic values allows animals to be ranked based on their aggregate genetic values in monetary terms, instead of biological values. Development and evaluation in monetary terms of breeding objectives is necessary to identify the most appropriate breeding objective for maximum productivity and economic efficiency. Economic return for a breeding program makes it possible to evaluate that breeding program objectively (Conington et al., 2001). The conventional selection index development is mainly summing the products of breeding values and economic values (Banga, 2009; Groen, 1990). The prerequisite for this method is to have similar traits in the breeding objective and selection criteria. In the approach suggested by Gibson (1995) and Schneeberger et al. (1992), the traits in the breeding objective can be different from the selection criteria, and therefore allows the use of indicator traits in selection indices. Application and evaluation of the approach of Gibson (1995) and Schneeberger et al. (1992) in practical breeding goal development has been limited, at least in the pig industry. Use of this approach may be beneficial to pig improvement programs as it will maximize profit, while minimizing the traits in the index. In addition, indicator traits that have been extensively used in selection programs may be objectively weighted during selection. The selection indices used by the South African breeding programs do not conform to the confines of economically relevant traits and selection index methodology as suggested by Bourdon (1998). Non-objective methods are used to derive index weights; hence the relationship between genetic improvement and profitability may be limited, which may restrain profit maximization by pig producers. Therefore, proper development and implementation of breeding objectives by objective methods is needed if the pig industry is to be productive, profitable and sustainable. The objectives of this study were to (i) develop breeding objectives that will maximize profit by using economic selection indices and (ii) evaluate the contribution made by indicator traits in South African Large White pig improvement programs.

2. Materials and methods 2.1. Performance testing and recording Pig performance testing was done at three testing stations, namely Irene, Elsenburg and Cedara, to evaluate production traits. The pigs were tested and slaughtered at these testing centers. Every year each member submitted 44 pigs (22 boars and 22 gilts) for testing. These pigs represented a minimum of five herd sires per breed or line, or 50% of the herd sires per breed or line. On arrival, the pigs were treated for internal and external parasites and quarantined, individually penned on solid concrete floors and fed until they commenced testing at 27 kg. During the test period, animals were individually housed and fed ad libitum using individual self-feeders and water was also available ad libitum from nipples. Daily feed intake was calculated as the difference between feed provided at the start of the day and the feed left in troughs at the end of the day. Backfat measurements were taken using a Backfat Scanner A100 probe at slaughter (86 kg), 6.5 cm from the midline between the second and the third last rib. Weighing during the test period was done weekly without any change in the feeding routine and performance testing. At the completion of the test, the pigs were fasted for 24 h before slaughter to empty stomach contents. Live weights were then measured, after which the pigs were electrically stunned with 250 V for 7–10 s before exsanguinations at the various abattoirs. Lean content (LEAN) was determined using a Hennesy Grading Probe. Animal ethics approval was obtained from the Agricultural Research Council Animal Ethics Committee. A state veterinarian or meat inspector performed the necessary inspection of the carcasses after slaughter. Dressing percentage (DRESS) was cold carcass weight expressed as a proportion of live weight at slaughter. 2.2. Biological traits affecting profits Sow productivity traits were number born alive (NBA), 21-day litter size (D21LS) and weight (D21LWT). Number born alive refers to the number of live piglets born to a particular sow in a particular farrowing. This affects the number of piglets that survive to weaning and the number of pigs marketed. Twenty-one-day litter size is the number of piglets from a particular sow's farrowing that reach 21 days of age, and the total weight of these piglets is D21LWT. Twenty-one-day litter weight affects post-weaning growth performance, ultimately affecting days to reach slaughter weight and the total costs incurred. The production traits included in the analyses were average daily gain (ADG), feed conversion ratio (FCR), age at slaughter (AGES), ultrasonic backfat thickness (BFAT), LEAN and DRESS. Ultrasonic backfat thickness indicates the level of carcass fatness and therefore reflects the leanness of a carcass. Average daily gain is the rate of body weight gain of a pig from birth to marketing and determines the number of days required to reach market weight (AGES), thus affecting production costs. Feed conversion ratio is the amount of feed consumed for a unit body weight gain, which should be reduced to improve

B. Dube et al. / Livestock Science 157 (2013) 9–19

Table 1 Summary statistics for the traits analyzed.

NBA (pigs) D21LS (pigs) D21LWT (kg) ADG (g/day) FCR (kg/kg) AGES (days) DRESS (%) LEAN (%) BFAT (mm)

N

Mean

Min

Max

SD

21,127 15,076 15,076 20,079 20,079 20,079 5406 5406 20,079

10.46 8.50 47.50 676.60 2.08 127.70 77.58 65.5 12.27

4.00 4.00 19.00 524.40 1.40 106.00 69.00 59.00 7.00

161.00 13.00 76.00 814.20 3.10 164.00 85.00 68.00 24.00

2.42 1.70 11.07 50.34 0.26 10.15 3.03 1.42 2.84

NBA – number born alive; D21LS – 21-day litter size; D21LWT – 21-day litter weight; ADG – average daily gain; FCR – feed conversion ratio; AGES – age at slaughter; DRESS – dressing percentage; LEAN – lean percentage; BFAT – backfat thickness.

Table 2 Distribution of records for sow productivity traits across the parities. Parity

Sows

Sires

Dams

Herds

HYS

1 2 3 4 5 6 7 8

7983 6281 4773 3542 2549 1792 1204 801

1189 1122 1024 924 797 658 503 394

3857 3344 2785 2193 1721 1293 907 625

29 28 28 26 26 25 24 22

330 326 316 294 269 243 218 190

profits. Thus, FCR is an important variable in reducing production costs (See et al., 1995). Even though LEAN should be improved to satisfy consumers, producers are also remunerated based on a unit weight of carcass, hence the need to improve carcass yield by improving DRESS. 2.3. Genetic parameter estimation The study on sow productivity and production traits was conducted on South African Large White pigs from across the country. Two data sets, one for sow productivity and one for production traits, were analyzed and there were no common pigs between the two data sets; hence there were no covariances between sow productivity and production traits. Data used in the analyses were obtained from the Integrated Registration and Genetic Information Systems (INTERGIS), while pedigree data were obtained from the Large White breed society. Table 1 shows the summary statistics for the traits analyzed. The data sets were edited to remove (i) records that were greater or less than three standard deviations from the mean and (ii) contemporary groups (HYS) with fewer than five animals and/or with fewer than two sires to ensure connectedness. Each HYS was created by concatenating herd, year and season of farrowing for sow productivity traits, while the HYS for production traits were created by concatenating herd, year and season of testing. The two seasons of farrowing or testing considered were summer (October– March) and winter (April–September). In the study on sow productivity, the traits were treated as traits of the dam to evaluate sow productivity. Table 2

11

shows the distribution of the sow productivity data across the eight parities for pigs that farrowed between 1990 and 2008. Data for production traits contained records for 20,079 pigs from 29 herds, performance-tested between 1990 and 2008, and records for 5406 carcasses from 20 herds, evaluated between 1993 and 2007. The data for growth traits had 1289 sires, 3216 dams and 340 contemporary groups, while there were 636 sires, 1289 dams and 178 contemporary groups in the data for carcass traits. Preliminary analyses were performed using the Mixed Procedure of SAS (2003) to determine the significance of the fixed effects and the conditional F-tests were implemented in the form of the ANOVA method. Production traits were adjusted for age at the beginning of the test. Random effects included in the analyses were determined using the log-likelihood ratio tests. The difference between the two log-likelihoods was multiplied by two and compared with Chi-squared values, with degrees of freedom equal to the difference between the original and the reduced model. Table 3 shows the fixed and random effects fitted in the analyses for the traits. Random effects included in the analyses of sow productivity were animal direct genetic and permanent environmental effects associated with each sow, as shown in Eq. (1). Eq. (2) shows the mixed model equation used in the analyses of production traits, where animal direct and maternal genetic effects were included as random effects. These genetic analyses were performed using REML procedures in ASREML (Gilmour et al., 2006). y ¼ XβþZ1 ua þWupe þe

ð1Þ

y ¼ XβþZ1 ua þZ2 um þe

ð2Þ

where y is the vector of observations, β is the vector of fixed effects, vectors of random effects consisted of random animal additive genetic (ua), permanent environmental (upe), maternal genetic (um) and residual (e) effects. Incidence matrices X, Z1, W and Z2 related fixed, direct Table 3 Fixed and random effects included in the genetic analyses of the traits. Trait

NBA D21LS D21LWT ADG FCR AGES DRESS LEAN BFAT

Fixed

Random

HYS

Parity

        

  

SEX

AFI

AFI2







  





 

 

Animal § § § § § § § § §

Maternal

PE § § §

§ § § § § §

HYS – herd, year and season of farrowing or testing; AFI – average daily feed intake; NBA – number born alive; D21LS – 21-day litter size; D21LWT – 21-day litter weight; ADG – average daily gain; FCR – feed conversion ratio; AGES – age at slaughter; DRESS – dressing percentage; LEAN – lean percentage; BFAT – backfat thickness; PE – permanent environmental effect of the sow;  – factor included and significant; § – random effect fitted.

12

B. Dube et al. / Livestock Science 157 (2013) 9–19

genetic, permanent environmental and maternal genetic effects, respectively to observations. 2.4. Base herd structure Stochastic simulations were performed for the dam and terminal sire lines to derive economic values for the sow productivity and production traits, respectively. In both simulations, maiden females were purchased at the age of 143 days and mated at the first estrus. Mating was by artificial insemination with purchased semen. Sows that failed to conceive after three consecutive services were culled. The non-pregnant sows were fed according to their nutrient requirements for growth. Pregnant sows were fed according to the nutrient requirements for pregnancy with the standard sow meal. After a 114-day gestation period, the sows farrowed and the numbers of male and female piglets were assumed to be the same. Male piglets were castrated at 10 days of age (castrates) and thus their nutrients requirements were assumed to be similar to those of gilts. The lactating sow was switched to a lactation diet and creep feeding commenced when the piglets were 13 days old until weaning at 28 days of age. After weaning, the sows were mated at the first estrus, while the piglets were fed the growers’ meal. Productive sows were allowed to have a maximum of eight litters. In the simulation for the dam line, growing and finishing commenced at 45 days of age for castrates until they reached 86 kg live weight, when they were marketed. Gilts started the growing and finishing phase at the age of 45 days until they reached 120 kg live weight, when they were marketed, while six percent of the gilts were kept as replacement females. For the terminal line, growing and finishing commenced at 45 days of age until the gilts and castrates reached 86 kg live weight, when they were all marketed. Culled-sows were marketed 45 days after weaning so that they could be fattened. Table 4 shows proportions of breeding sows, sows culled, live weights of cull-sows and litter sizes per parity. Input parameters for the production system are shown in Table 5 (South African Pork Producers Organization (SAPPO), 2008). 2.5. Economic value derivation A hypothetical 100-sow-herd was used for convenience of calculations. Five pig categories were distinguished according

Table 4 Herd composition of the pig population.

27.6 21.7 16.5 12.2 8.8 6.2 4.2 2.8

15 15 15 30 30 30 50 100

2.6. Selection index development The fact that the importance of each trait in a pig enterprise varies according to its relationship with profit suggests the need to use economic values to weight EBVs in multi-trait selection programs. Where the EBVs are simply weighted by their respective economic values, the original selection indices I were written simply as (Gibson, 1995) n

I ¼ ∑ ai EBVi

Parity Proportion of Sows culled after Weight of each Litter breeding sows (%) weaning (%) cull-sow (kg) size 1 2 3 4 5 6 7 8

to age: (1) piglets (0–28 days old); (2) weanlings (29–44 days old); (3) gilts (≥45 days and ≤120 kg); (4) castrates (≥45 days and ≤86 kg); (5) breeding sows (≥120 kg). Total annual profit of the herd was derived as the difference between revenues and costs of the system. All costs and prices were expressed in South African Rand (ZAR) per sow per generation, which is assumed to be 3 years with 2.5 litters per year. The inputs for the production system were feed (creep, grower and finisher, and sow meal), management (labor, spraying or dipping and veterinary services), marketing (transport of live pigs) and fixed costs (housing, electricity, stationery and interests). The outputs were the revenues from sale of porkers, maiden gilts and cull-sows. Supply of labor by the farmer was set to be fixed per pig per year and was considered equal for all pig categories except for replacement stock. Replacement stock was assumed to need less care than the young stock and breeding pigs, thus it was considered to require half the amount of labor per pig. Veterinary care was assumed to be optimal and therefore, reasonable average costs have been used. Other costs not related to herd size were included in the fixed costs, shown in Table 6 (SAPPO, 2008). Mating was done by artificial insemination and, cost of semen and insemination costs were included in the veterinary costs. The profit functions used are shown in Supplementary file. Economic values for sow productivity traits, ADG, FCR, DRESS and AGES were derived through the partial budget approach. Partial differentiation of the profit function was used to derive economic values for LEAN and BFAT. Change in profit was calculated by evaluating herd profit (Tf1) numerically for each trait, then evaluating it after incrementing by one unit of that trait (thus obtaining Tf2), and taking the difference Tf2−Tf1 (Ponzoni, 1992). The profit difference was then multiplied by the genetic standard deviation to produce an economic value. Relative economic values were computed by setting NBA and BFAT as base traits against which comparisons for sow productivity and production traits would be made, respectively.

138 147 170 188 196 210 220 220

9.71 10.41 11.07 11.24 11.24 11.10 11.00 10.76

i¼1

where i is the ith trait, n is the number of traits, ai is the economic value of the ith trait, and EBVi is the EBV of the ith trait. Possible breeding objectives were constructed, including their possible corresponding selection criteria to be used in selection programs. Average daily gain shows the rate of body weight gain and this is reflected in the number of days a pig takes to reach slaughter weight. Age at slaughter was therefore included either as an

B. Dube et al. / Livestock Science 157 (2013) 9–19

13

Table 5 Input parameters for the production systems. Source: SAPPO (2008). Parameter (unit)

Value

Parameter (unit)

Number of sows 100 Farrowing percentage (%) 90 Gestation period (days) 114 Maximal number of farrowings 8 Conception rate (%) 90 Number of litters per sow per year 2.5 Preweaning mortality (%) 10 Post-weaning mortality (%) 5 Culling rate (%) 20 Replacement rate (%) 25 Gestation maternal weight gain (kg) 20 Production percentage (%) 95 Age of gilts at purchase (days) 143 Purchase price of gilts (ZAR/pig) 1500.00 Interval from purchase to culling of gilts not becoming pregnant (days) 63 Interval from purchase to 1st mating of gilts (days) 30 Maximum estrus cycles after weaning 3 Interval from the last weaning to culling of sows (days) 45

Table 6 Fixed costs of production for the production systems. Source: SAPPO (2008). Cost

Amount (ZAR) per herd

Administration fees Bank charges Cell phone costs Depreciation Electricity and water Fuel and oil Interest paid – long term loan Printing and stationery Protective clothing Repairs and maintenance Salaries Veterinary costs Wages

12,000.00 3000.00 3000.00 105,075.00 18,000.00 24,000.00 246,042.00 1800.00 3000.00 9,000.00 96,000.00 36,000.00 36,000.00

indicator trait for ADG or as a breeding objective trait. This was to establish which of the two traits (ADG or AGES) could be recommended for selection programs to improve growth performance. Lean content is measured in carcasses, which requires the animal to be slaughtered and may also result in few records used in genetic selection programs. Typically, breeding programs select against backfat thickness, measured on live animals, to improve carcass lean content (Nguyen and McPhee, 2005). These breeding programs, however, do not specify whether backfat thickness is included as a breeding objective trait or an indicator trait for lean content. In this study, backfat thickness was included either as a breeding objective trait or an indicator trait for LEAN. Inclusion of AGES and BFAT as breeding objective traits was made possible by the availability of their economic values. This caused changes in the original breeding objectives, resulting in different breeding objectives to be evaluated. Feed conversion ratio is not easily and accurately measured since it involves meticulous recording of feed intake. The value of including FCR in the selection criteria was determined by removing it from the selection criteria, while including it in the breeding

Value

Interval from weaning to 1st insemination (days) 7 Proportion of purchased gilts culled without farrowing (%) 15 Live weight of culled gilts (kg) 130 Lactation length (days) 28 Number born alive (pigs) 10.59 Gilts as percentage of NBA 50 Piglet birth weight (kg) 1.5 Mature pig weight (kg) 250 Pre-weaning weight gain (g/day) 183 Post-weaning weight gain (g/day) 713.84 Post-weaning feed conversion ratio 2.08 Lean percentage (%) 65.5 Backfat thickness (mm) 12.27 Sale weight for porkers (kg) 86.00 Marketing price for porkers (ZAR/kg live weight) 13.00 Marketing price for porkers (ZAR/kg carcass weight) 15.00 Transport costs (ZAR/km) 10.00

objective. As a result, the traits in the breeding objective were different from those in the selection criteria. In situations where the breeding objective was different from the selection criteria, the index proposed by Gibson (1995) and Schneeberger et al. (1992) was used, where the index weights (bI) were bI ¼ C−1 I CIH a where CI is the genetic variance–covariance matrix among the selection criteria, CIH is a matrix of genetic covariances between breeding objective traits and traits in the selection criteria, and a is the vector of economic values. Genetic variances and covariances among the traits are shown in Table 7. The index weights would be used as weights for EBVs for the selection criteria during genetic selection. Due to the different combinations of selection criteria and breeding objectives, three combinations of breeding objectives and selection criteria were developed for sow productivity traits, while there were 12 combinations for production traits. The development of several indices enabled evaluations and comparisons among indices to determine the most appropriate selection index and its corresponding breeding objective. Evaluation of breeding objectives was done by calculating genetic response to selection and economic return. The methodology to predict genetic responses to index selection when breeding objective traits were similar to selection criteria was done as described by Groen (1990). Therefore, the vector of predicted responses to selection for breeding objective traits (RH) was a′CH RH ¼ i pffiffiffiffiffiffiffiffiffiffiffiffiffi a′CH a where i is the selection intensity set at 1.4 (best 20% using truncation selection), which is implemented after the 6% replacement gilts have been incorporated into the herd. The assumption is that the replacement gilts are genetically superior and fall within the best 20% of the herd. CH is the genetic variance–covariance matrix for breeding

14

B. Dube et al. / Livestock Science 157 (2013) 9–19

Table 7 Genetic variances (diagonal), genetic covariances (above diagonal) and genetic correlations (below diagonal) for the traits analyzed.

NBA D21LS D21LWT ADG FCR AGES DRESS LEAN BFAT

NBA

D21LS

D21LWT

0.31 0.88 70.04 0.32 7 0.07

0.20 0.09 0.78 7 0.05

0.22 1.35 8.81

ADG

FCR

AGES

DRESS

LEAN

BFAT

129.9 −0.617 0.03 −0.927 0.03 −0.027 0.26 0.09 7 0.20 −0.017 0.18

−1.71 0.01 −0.747 0.09 −0.0770.02 −0.277 0.02 0.357 0.15

−77.92 0.68 17.36 0.1870.26 −0.27 70.20 0.29 70.21

1.14 −0.02 0.60 0.67 0.08 70.24 −0.107 0.23

0.22 −0.01 −0.81 0.05 0.49 −0.9170.03

−0.19 0.06 1.71 −0.12 −0.89 1.98

NBA – number born alive; D21LS – 21-day litter size; D21LWT – 21-day litter weight; ADG – average daily gain; FCR – feed conversion ratio; AGES – age at slaughter; DRESS – dressing percentage; BFAT – backfat thickness; LEAN – lean content.

objective traits. Genetic change was expressed per year, where the generation interval was assumed to be three years. When breeding objective traits and selection criteria were different, the method developed by Groen (1990), which was consistent with the method proposed by Schneeberger et al. (1992), was adopted to calculate genetic change. Thus, the responses to selection for breeding objective traits (RH) were computed as b′CIH R H ¼ i pffiffiffiffiffiffiffiffiffiffiffiffiffiffi b′CIH a Total economic return (ER) in ZAR from selection using the index was calculated by multiplying the vector of economic values by the vector of responses to selection for breeding objective traits (Wolfova et al., 2007). ER ¼ a′RH

2.7. Sensitivity analyses Due to the changing nature of feed and marketing prices, economic values have the tendency to change (Kosgey et al., 2003). Sensitivity analyses were conducted to determine the effect of changes in economic values on economic return. The sensitivity analyses were conducted in two ways. Firstly, the economic values for all traits were simultaneously increased successively by 10% starting from 10% to 40%, and then decreased by the same amounts. Then, the economic value for each trait was increased successively by 10% starting from 10% to 40%, while keeping other traits’ economic values constant, and then reduced by the same amounts. Responses to selection for breeding objective traits and economic returns were evaluated in the sensitivity analyses. The evaluations were done on the recommended selection indices and breeding objectives. 3. Results and discussion 3.1. Genetic parameters Table 7 shows the genetic covariances and correlations among the traits analyzed, while the heritability estimates, repeatability estimates and maternal genetic effects for these

Table 8 Heritability estimates, repeatability estimates and maternal genetic effects for the traits analyzed. Trait

h2

r

NBA D21LS D21LWT ADG AGES DRESS LEAN BFAT

0.0770.01 0.03 70.01 0.06 70.01 0.25 70.05 0.26 70.08 0.15 70.07 0.32 70.08 0.50 70.04

0.157 0.01 0.117 0.01 0.127 0.01

m2

ram

0.117 0.04 0.137 0.05 0.107 0.08 0.167 0.08 0.107 0.02

−0.477 0.17 −0.417 0.19 −0.45 7 0.23 −0.38 7 0.15 −0.62 7 0.06

NBA – number born alive; D21LS – 21-day litter size; D21LWT – 21-day litter weight; ADG – average daily gain; FCR – feed conversion ratio; AGES – age at slaughter; DRESS – dressing percentage; BFAT – backfat thickness; LEAN – lean content; h2 – heritability estimate; r – repeatability estimate; m2 – maternal genetic effects; ram – correlation between direct animal genetic and maternal genetic effects.

traits are shown in Table 8. Sow productivity traits had very low heritability estimates, suggesting that genetic improvement may be slow when genetic selection is applied on them. These estimates are consistent with most results in literature reported for other breeds (Chimonyo et al., 2006; Roehe et al., 2009). Litter size has low heritability probably because it is a composite trait (Bennett and Leymaster, 1990; Haley and Lee, 1992). Repeatability estimates for the sow productivity traits were also generally very low. Fernández et al. (2008) regarded mothering ability measured by D21LWT as a composite trait, combining litter size and average piglet weight at 21 days, which depend on the number of piglets born, as well as the piglets’ pre-weaning survival and growth. These results indicate that the sow's performance is unlikely to be repeated during the production period. This may be because physiological development of reproductive organs differs with parity (Oh et al., 2006). Therefore, sow performance during the first parity may not be used to predict its performance in future parities in these traits. The heritability estimates for production traits were moderate to high and are comparable to literature estimates (Chimonyo and Dzama, 2007; Nguyen and McPhee, 2005). These results show that there is substantial genetic variation in this population; hence genetic improvement may be achieved if genetic selection is applied on these traits. The contributions of maternal genetic effects to the

B. Dube et al. / Livestock Science 157 (2013) 9–19

phenotypic variances were less than the corresponding estimates for direct heritabilities. Mohuiddin (1993) pointed out that maternal heritabilities tend to be lower than direct heritabilities, indicating a greater genetic influence of the animal than its dam for the trait. The maternal genetic effects were slightly higher than those reported by Chimonyo and Dzama (2007) and Chen et al. (2002), which contributed less than 10% to the phenotypic variation of most traits. The genetic correlations between direct and maternal effects were all negative, consistent with previous reports (Chen et al., 2002; Chimonyo and Dzama, 2007). The antagonism suggests that both direct and maternal components should be taken into account during genetic selection to achieve optimum genetic progress (Johnson et al., 2002). The genetic correlations among sow productivity traits ranged from 0.32 70.07 between NBA and D21LWT to 0.8870.04 between NBA and D21LS. For production traits, the genetic correlations ranged from no correlation between ADG and BFAT to −0.9270.03 between ADG and AGES. High genetic correlations suggest that selection for one trait can result in a correlated response in the other. On the other hand, low genetic correlations suggest that both traits should be included in the selection program. 3.2. Economic values Economic values were computed at 95% production by taking the difference between revenues and costs (Ponzoni, 1988). The economic values and the genetic standard deviations for the traits analyzed depended on population means. The profit differences, economic values and relative economic values for the traits are shown in Table 9. Contrary to expectations, the economic value for NBA was higher than that for D21LS. This is attributed to the lower genetic variation for D21LS compared to that for NBA, despite the higher profit difference for D21LS. When there is less genetic variation there will be less room for genetic improvement and this will lower the economic value associated with it.

Table 9 Profit differences, economic values and relative economic values for the traits analyzed. Trait Sow productivity NBA D21LS D21LWT Production ADG FCR AGES DRESS LEAN BFAT

ΔP (ZAR)

EV (ZAR)

REV

110.02 166.93 74.93

61.51 50.02 222.39

1.00 0.81 3.62

2.93 −233.83 −16.46 7.07 6.70 −1.05

33.39 −21.04 −68.64 5.80 4.69 −1.48

22.52 −14.74 −46.07 3.91 3.17 −1.00

ΔP – profit difference; EV – economic value; REV – relative economic value; NBA – number born alive; D21LS – 21-day litter size; D21LWT – 21-day litter weight; ADG – average daily gain; FCR – feed conversion ratio; AGES – age at slaughter; DRESS – dressing percentage; BFAT – backfat thickness; LEAN – lean content.

15

The study highlights the importance and contribution of genetic variation to economic values. Twenty-one-day litter weight has an economic value of ZAR 222.39 per kg genetic change. This kg change is the aggregate change for the entire litter, which has implications on the post-weaning growth performance, survival and age at slaughter for each pig. The costs and income associated with the D21LWT kg change produce the observed economic value. From an animal improvement point of view, only the range around the actual population mean is interesting for short-to-medium-term considerations (Von Rohr et al., 1999). This is probably because the economic environment is constantly changing, which requires the continual updating of economic values. Since the economic environment varies from one environment to another and from herd to herd, genetic selection programs can be custom-made to suit a particular environment or herd. In this study, the economic value for FCR was estimated to determine by how much improvement of feed utilization will improve profit. Economic weights for feed intake, daily gain and lean % were estimated by Hermesch et al. (2003), and were consistent with the findings of the current study. Houska et al. (2004) expressed economic values per purchased gilt and year, per fattened animal and per kg of slaughter weight. These economic values indicate the direction and emphases of genetic selection to be applied on the respective traits. Relative economic values indicate how much additional profit can be earned by making an improvement of one standard deviation in one trait relative to NBA in sow productivity traits or BFAT in production traits.

3.3. Selection indices Different breeding objectives were evaluated to determine the breeding objective that would optimize genetic improvement of all the traits and maximize profits. Selection indices provide the optimal selection emphasis on different traits, based on their relative economic importance (Conington et al., 2001). These selection indices depend highly on genetic parameters and economic values, suggesting that estimation of these components should be done with utmost accuracy. MacNeil et al. (1997) noted that small errors in economic values or genetic parameters may only lead to minor losses in efficiency of selection. Some studies have used selection indices based on BLUP, where EBVs were weighted by economic values (e.g., Wolfova et al., 2007). The current study was based on this method of index development. Indices 1 and 4 (Table 10) were based on the complete breeding objectives, which were also the selection criteria; hereafter referred to as complete indices. This study determined whether these indices were the most appropriate for the improvement of sow productivity and production traits. It also evaluated the prospects of utilizing alternative breeding objectives and selection indices, including the use of indicator traits. Genetic responses and economic returns for sow productivity traits are shown in Table 10. The genetic responses for D21LS and D21LWT were the lowest in the complete index and the highest in the index that was composed of NBA and

B. Dube et al. / Livestock Science 157 (2013) 9–19

16

Table 10 Breeding objectives, selection indices and genetic responses to selection and economic returns. Index

1 2 3

Selection criteria

NBA, D21LS, D21LWT D21LS, D21LWT NBA, D21LWT

Breeding objective and genetic responses NBA

D21LS

D21LWT

0.053 0.030 0.053

0.209 0.215 0.222

1.345 1.385 1.381

ADG 4 5 6 7 8 9 10 11 12 13 14 15

ADG, FCR, DRESS, LEAN ADG, DRESS, LEAN AGES, DRESS, LEAN AGES, DRESS, LEAN ADG, FCR, DRESS, BFAT ADG, DRESS, BFAT AGES, DRESS, BFAT AGES, DRESS, BFAT ADG, FCR, DRESS, BFAT ADG, DRESS, BFAT AGES, DRESS, BFAT AGES, DRESS, BFAT

ER (ZAR)

AGES

5.32 5.32 9.41 −1.94 5.32 5.32 9.50 −1.94 5.32 5.32 9.50 −1.94

FCR

DRESS

LEAN

−0.07 −0.07 −0.08 −0.08 −0.07 −0.07 −0.08 −0.08 −0.07 −0.07 −0.08 −0.08

0.05 0.05 0.03 −0.06 0.05 0.05 0.03 −0.06 0.05 0.05 0.03 −0.06

0.01 0.01 0.01 0.09

312.81 320.69 321.58 BFAT

−0.01 −0.01 −0.01 −0.20 0.01 0.01 0.01 0.09

179.19 179.19 315.64 134.29 179.19 179.16 318.50 134.14 179.19 179.19 318.52 134.29

NBA – number born alive; D21LS – 21-day litter size; D21LWT – 21-day litter weight; ADG – average daily gain; FCR – feed conversion ratio; AGES – age at slaughter; DRESS – dressing percentage; BFAT – backfat thickness; LEAN – lean content; ER – economic return.

D21LWT. Removing NBA from the index and including it in the breeding objective reduced its genetic response to selection. The complete index produced the lowest economic return. Excluding 21-day litter size from the index and including it in the breeding objective produced the highest economic return, which was 2.8% higher than in the complete index. There was a 2.5% improvement in economic return when the index that consisted of D21LS and D21LWT was used. Thus, for sow productivity traits, economic return may not be maximized by simply weighting EBVs for NBA, D21LS and D21LWT with their respective economic values. Removing D21LS from the index may also be justified by the very high genetic correlations it has with NBA and D21LWT (0.8870.04 and 0.7870.05, respectively). The recommended index from the current study is different from the USA Swine Testing and Genetic Evaluation Systems (STAGES) sow productivity index (Schinckel and Einstein, 1999). In the sow productivity index for STAGES, number born alive, number weaned and 21-day litter weight are weighted by their respective economic values. Table 10 contains the genetic responses to selection and economic returns for production traits. Several breeding programs select directly for average daily gain to improve growth performance (e.g., Suzuki et al., 2005), including the South African pig breeding program (Visser, 2004). The current study determined the best way to improve growth performance. This was made possible by the availability of genetic parameters and economic values for ADG and AGES. Including AGES as a breeding objective trait may have unfavorable responses in carcass yield as evidenced by the decreasing genetic gains for DRESS. Economic return was also lower when AGES was included as a breeding objective trait. However, including AGES as a breeding objective trait resulted in genetic improvements in FCR, LEAN and BFAT. Therefore, when improvement of feed utilization and carcass quality

are a priority in this population, AGES can be included as a breeding objective trait. When AGES was included as an indicator trait for ADG, there was a greater genetic response in ADG to selection than from direct inclusion of ADG into the index. This may suggest that slower genetic progress may be expected in growth rate when selection for ADG is done by including it in an index. Conversely, AGES, which is determined once at marketing, may be included as an indicator trait for ADG to achieve faster genetic progress. Economic return was also higher when AGES was included as an indicator trait for ADG. This increased economic return may be attributed to double-counting of growth in ADG and AGES. Use of AGES is consistent with the terminal sire index used by the STAGES, where days to slaughter are included in the index, instead of average daily gain (Schinckel and Einstein, 1999). In the STAGES however, age at slaughter is included as a breeding objective trait. Breeding programs prefer to improve carcass leanness by reducing backfat thickness measured on live animals (Nguyen and McPhee, 2005; Suzuki et al., 2005). Such programs however do not specify how backfat thickness is included in those selection programs. In the STAGES terminal sire index, backfat thickness is included as a breeding goal trait (Schinckel and Einstein, 1999). The South African pricing system makes use of the two traits (BFAT and LEAN), such that during marketing, either of the two traits is evaluated, depending on whether it is a live pig or a carcass being marketed. In this study, economic values and genetic parameters for the two traits were estimated. This study established the most appropriate way to include BFAT in pig breeding programs, since breeding programs for this population also do not specify how BFAT is included. Including BFAT as an indicator trait for LEAN produced results similar to including LEAN in the

B. Dube et al. / Livestock Science 157 (2013) 9–19

17

Table 11 Sensitivity analyses of genetic responses to selection and economic returns to changes in economic values. %Δ

NBA (pigs)

D21LS (pigs)

D21LWT (kg)

ΔERS (ZAR)

ADG (g/day)

FCR

DRESS (%)

LEAN (kg)

ΔERP (ZAR)

+10 −10 +20 −20 +30 −30 +40 −40

0.005 0.000 0.008 −0.003 0.010 −0.006 0.012 −0.009

0.032 −0.106 0.097 −0.181 0.158 −0.261 0.218 −0.347

0.007 −0.015 0.018 −0.027 0.028 −0.040 0.037 −0.053

7.29 −22.95 21.37 −39.34 34.88 −56.78 47.88 −75.53

0.463 −4.87 0.906 −1.003 1.331 −1.551 1.740 −2.782

–0.003 0.004 –0.007 0.009 –0.010 0.013 –0.014 0.023

0.002 −0.001 0.003 −0.003 0.004 −0.005 0.005 −0.008

0.001 0.000 0.001 0.000 0.002 −0.001 0.002 −0.002

15.55 −16.34 30.40 −33.62 44.65 −52.02 58.36 −93.29

%Δ – percentage change; NBA – number born alive; D21LS – 21-day litter size; D21LWT – 21-day litter weight; ΔERS – change in economic return for sow productivity; ADG – average daily gain; DRESS – dressing percentage; LEAN – lean content; FCR – feed conversion ratio; ΔERP – change in economic return for production traits.

index and breeding objective (Indices 4–7 and 12–15). On the other hand, including BFAT as a breeding objective trait resulted in lower economic returns (Indices 8–11). Therefore, the best way to include BFAT in breeding programs for this population is to include it as an indicator trait for LEAN. Improving ADG and BFAT may result in increased daily maintenance costs due to the increased nutrient requirements of the growing pig (Cleveland et al., 1983). This kind of selection is expected to result in increased mature size and increase maintenance requirements of the breeding herd. In the current study, however, inclusion of FCR in the breeding objective may mitigate against such unfavorable responses. The STAGES terminal sire index includes FCR in the breeding objective and excludes it from the index (Schinckel et al., 1998). A similar approach was performed in this study and there was no effect on responses to selection and economic returns from the recommended breeding objective. Feed conversion ratio may thus be improved by selecting for improved growth performance, carcass yield and quality. These results are consistent with previous reports where, improvements in feed conversion ratio due to selection for increased average daily gain and reduced backfat thickness were reported (Jungst et al., 1981; Vangen, 1980). The most appropriate breeding objective for production traits consisted of ADG, FCR, DRESS and LEAN with a corresponding index consisting of AGES, BFAT and DRESS (Index 14). Age at slaughter and BFAT were included as indicator traits for ADG and LEAN, respectively. The favorable results were, however, obtained from an index that had positive weights for BFAT. Producers may be reluctant to weight BFAT positively and prefer indices that place negative weights on BFAT. Under such circumstances, the producers may utilize the index that consists of ADG, DRESS and BFAT (Index 9) with the same breeding objective, despite the lower economic returns expected. In this case, BFAT should still be included as an indicator trait for LEAN. Gibson (1995) suggested the use of a restricted selection index in the case where producers are not willing to place positive weights on backfat thickness. The recommended breeding objectives and indices are different from those currently utilized by the South African pig industry. The index weights used by the South African pig industry are derived using non-objective methods. In addition, the

indices do not conform to the rules of economically relevant traits and selection index methodology. In the indices used by the South African pig industry, LEAN and BFAT are in the same index, where BFAT can be considered an indicator trait for LEAN; hence it cannot be in the same index. The same applies to ADG and AGES, where AGES can be an indicator trait for ADG. Banga (2009) pointed out that if the economically relevant trait is included in the index, then its indicator trait has no value. This study showed the antagonism between direct and maternal genetic effects, which may reduce direct genetic progress. Genetic responses to selection computed in this study did not consider maternal genetic effects; as a result they may be overestimated. Therefore, to counter the maternal effects of reducing performance, genetic responses to selection should be adjusted for maternal genetic effects. The method for calculating responses to selection when the traits in the breeding objective are different from those in the index was adopted from Groen (1990) and is consistent with the method suggested by Schneeberger et al. (1992). In the approach by Schneeberger et al. (1992), a matrix of EBVs for the selection criteria is first constructed. Adopting the method proposed by Groen (1990) may be a more direct method of computing responses to selection from indices based on BLUP as it bypasses the computation of the EBV variance–covariance matrix. This study highlights the contribution made by indirect selection, through the use of indicator traits for economically relevant traits, to herd improvement. 3.4. Sensitivity analyses Since pig enterprises aim to maximize profit, their selection programs, which ensure that this is achieved, depend on economic values. The economic climate is constantly changing due to various economic factors, which cause economic values to also change. For breeding programs to place the correct emphases on traits during selection, the changes in economic values should be considered. The sensitivity of genetic changes and economic returns to changes in economic values were determined for the recommended selection indices for sow productivity traits (Index 3) and production traits (Index 14). Genetic changes and economic returns for sow productivity and production traits were responsive to changes in

18

B. Dube et al. / Livestock Science 157 (2013) 9–19

economic values. The results for genetic changes and economic returns when economic values for traits were changed simultaneously are shown in Table 11. For sow productivity, the greatest and lowest changes occurred in D21LS and NBA, respectively. In a study by Houska et al. (2004) on Czech Republican pigs, the economic values for number born alive and litter size were sensitive to changing economic circumstances. This underlines the impact of the economic environment on the reproductive performance of commercial sows. When economic values were changed for all traits simultaneously, the greatest changes for production traits were in ADG, while the lowest were in LEAN. Generally, genetic changes and economic return increased when economic values increased and decreased when economic values decreased. The changes were however not symmetrical, as the magnitudes of the decrease were always greater than the magnitudes of the increase. This may suggest that when economic values decrease, the losses in genetic progress and economic return may be greater than the gains obtained when economic values increase by the same magnitude. When economic values were changed for one trait at a time, genetic change for LEAN responded to changes in economic values for all the traits, while genetic change for FCR only responded to changes in the economic value for ADG. The genetic change for ADG was only sensitive when economic values for ADG and FCR changed. This may be consistent with the study conducted by Houska et al. (2004), which showed that the economic value for average daily gain is sensitive to changing economic parameters. Genetic changes for all the traits were sensitive to changes in the economic value for ADG. The greatest change in economic return occurred when the economic value for ADG changed. On the other hand, the lowest was observed when there was a change in the economic value for LEAN. This may be attributed to the magnitudes of these traits’ economic values and genetic variances. Conington et al., 2001 attributed changes in genetic progress to changes in economic values in the UK sheep. This indicates that changes in economic parameters that modify economic values for breeding goal traits may alter genetic progress. Breeders should therefore constantly recalculate economic values to place the appropriate emphases on traits during genetic selection. Economic values can be recalculated when there is a significant reduction in the change in economic returns in the pig industry. The results from this study were obtained from a simulation where pigs were individually fed ad libitum. This is to ensure that there is no nutritional deficiency, thus allowing pigs to perform according to their genetic potentials. In a commercial system however, pigs are group-fed with the same objective of maximum genetic utilization of feed. The observations from this study can therefore be extrapolated to a commercial system. 4. Conclusions Economic evaluation of genetic selection programs has been made possible by the availability of economic values and genetic parameters. Application of indirect selection

by using indicator traits may be useful in genetic improvement programs for this population. Improvement of sow productivity may be achieved by selecting on an index that consists of NBA and D21LWT. Growth performance, and carcass yield and quality may be improved by selecting on an index comprising of AGES, DRESS and BFAT. The EBVs of the index traits would be weighted by weights generated using the genetic parameters and economic values. Highest economic returns may be expected when these indices are utilized. Inclusion of AGES and BFAT in the selection program improves growth rate and carcass leanness, respectively. Lower economic returns may be expected when AGES is included in genetic selection programs as a breeding objective trait. This study highlights the importance of specifying how a trait is included in the breeding program (breeding objective trait or indicator trait). Economic values for breeding goal traits should be recalculated periodically, as genetic progress and economic returns change with changes in economic values. Conflict of interest statement The data for the study were obtained from the Agricultural Research Council (ARC), South African Pork Producers Organization (SAPPO) and South African Meat Industry Company (SAMIC).

Acknowledgments The data for genetic analyses were obtained from the Agricultural Research Council (ARC). Information on production systems was from the South African Pork Producers Organization (SAPPO), while the pricing information was from the South African Meat Industry Company (SAMIC). Appendix A. Supporting information Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j. livsci.2013.06.018.

References Banga, C.B., 2009. The Development of Breeding Objectives for Holstein and Jersey Cattle in South Africa. University of Free State, Bloemfontein, South Africa (Ph.D. thesis). Bennett, G.L., Leymaster, K.A., 1990. Genetic implications of a simulation model of litter size in swine based on ovulation rate, potential embryonic viability and uterine capacity: II. Simulated selection. J. Anim. Sci. 68, 980–986. Bourdon, R.M., 1998. Shortcomings of current genetic evaluation systems. J. Anim. Sci. 76, 2308–2323. Chen, P., Baas, T.J., Mabry, J.W., Dekkers, J.C.M., Koehler, K.J., 2002. Genetic parameters and trends for lean growth rate and its components in U. S. Yorkshire, Duroc, Hampshire, and Landrace pigs. J. Anim. Sci. 80, 2062–2070. Chimonyo, M., Dzama, K., 2007. Estimation of genetic parameters for growth performance and carcass traits in Mukota pigs. Animal 1 (3), 317–323. Chimonyo, M., Dzama, K., Bhebhe, E., 2006. Genetic determination of individual birth weight, litter weight and litter size in Mukota pigs

B. Dube et al. / Livestock Science 157 (2013) 9–19

parameters for growth performance and carcass traits in Mukota pigs. Livest. Sci. 105, 69–77. Cleveland, E.R., Johnson, R.K., Mandigo, R.W., Peo Jr., E.R., 1983. Index Selection and Feed Intake Restriction in Swine. II. Effect on Energy Utilization. Faculty Papers and Publications in Animal Science. University of Nebraska, Lincoln570–578. Conington, J., Bishop, S.C., Grundy, B., Waterhouse, A., Simm, G., 2001. Multi-trait selection indexes for sustainable UK hill sheep production. Anim. Sci. 73, 413–423. Fernández, A., Rodrigáñez, J., Zuzúarregui, J., Rodríguez, M.C., Silió, L., 2008. Genetic parameters for litter size and weight at different parities in Iberian pigs. Span. J. Agric. Res. 6, 98–106. Gibson, J.P., 1995. An Introduction to the Design and Economics of Animal Breeding Strategies. Course Notes. Prague-Uhrineves, 7–16 September. Gilmour, A.R., Gogel, B.J., Cullis, B.R., Thompson, R., 2006. ASReml User Guide Release 3.0. VSN International Ltd, Hemel Hempstead, HP1 1 ES, UK. Groen, A.F., 1990. Influences of production circumstances on the economic revenue of cattle breeding programmes. Anim. Prod. 51, 469–480. Haley, C.S., Lee, G.J., 1992. Genetic factors contributing to variation in litter size in British Large White gilts. Livest. Prod. Sci. 30, 99–113. Hazel, L.N., 1943. The genetic basis for constructing selection indexes. Genetics 28, 476–490. Hermesch, S., Kanis, E., Eissen, J.J., 2003. Economic weights for feed intake in the growing pig derived from a growth model and an economic model. J. Anim. Sci. 81, 895–903. Houska, L., Wolfova, M., Fiedler, L., 2004. Economic weights for production and reproduction traits of pigs in the Czech Republic. Livest. Prod. Sci. 85, 209–221. Johnson, Z.B., Chewning, J.J., Nugent 3rd, R.A., 2002. Maternal effects on traits measured during postweaning performance test of swine from four breeds. J. Anim. Sci. 80, 1470–1477. Jungst, S.B., Christian, L.L., Kuhlers, D.L., 1981. Response to selection for feed efficiency in individually fed Yorkshire boars. J. Anim. Sci. 53, 323–332. Kosgey, I.S., Arendonk, J.A.M., Baker, R.L., 2003. Economic values for traits of meat sheep in medium to high production potential areas of the tropics. Small Ruminant Res. 50, 187–202. MacNeil, M.D., Nugent, R.A., Snelling, W.M., 1997. Breeding for Profit: an introduction to selection index concepts. In: Proceedings of the Range Beef Cow Symposium XV December 9, 10 and 11, 1997. Rapid City, South Dakota. Mohuiddin, G., 1993. Estimates of genetic and phenotypic parameters of some performance traits in beef cattle. Anim. Breed. Abstr. 61, 495–522. Nguyen, N.H., McPhee, C.P., 2005. Genetic parameters and responses of performance and body composition traits in pigs selected for high

19

and low growth rate on a fixed ration over a set time. Genet. Sel. Evol. 37, 199–213. Oh, S.H., Lee, D.H., See, M.T., 2006. Estimation of genetic parameters for reproductive traits between first and later parities in pigs. Asian– Aust. J. Anim. Sci. 19, 7–12. Olesen, I., Groen, A.F., Gjerde, B., 2000. Definition of animal breeding goals for sustainable production systems. J. Anim. Sci. 78, 570–582. Ponzoni, R.W., 1992. Which trait for genetic improvement of beef cattle reproduction: calving ease or calving day? J. Anim. Breed. Genet. 109, 119–128. Ponzoni, RW., 1988. The derivation of economic values combining income and expense in a different way: an example with Australian Merino sheep. J. Anim. Breed. Genet. 105, 143–152. Roehe, R., Shrestha, N.P., Mekkawya, W., Baxter, E.M., Knap, P.W., Smurthwaite, K.M., Jarvis, S., Lawrence, A.B., Edwards, S.A., 2009. Genetic analyses of piglet survival and individual birth weight on first generation data of a selection experiment for piglet survival under outdoor conditions. Livest. Sci. 121, 173–181. SAS, 2003. SAS User's Guide. SAS Institute Inc., Cary, NC, USA. Schinckel, A.P., Einstein, M.E., 1999. STAGES Index Changes for 1999. Purdue University Swine Day, August 31. Schinckel, A.P., Lofgren, D.L., Einstein, M.E., 1998. Recent STAGES Index Changes. Purdue University Swine Day, September 3. Schneeberger, M., Barwick, S.A., Crow, G.H., Hammond, K., 1992. Economic indices using breeding values predicted by BLUP. J. Anim. Breed. Genet. 109, 181–187. See, T., Zering, K., Robison, O.W., 1995. Economic Value of Pork Quality Traits. 〈http://www.nsif.com/conferences/1995/evnsif.htm〉. (Retrieved 14.04.10). South African Pork Producers Organization, 2008. Research and Development Plan for the Pork Industry in South Africa. RMRD Planning Committee (R & D) Pork. 2nd Version June 2008. Pretoria, South Africa. Suzuki, K., Irie, M., Kadowaki, H., Shibata, T., Kumagai, M., Nishida, A., 2005. Genetic parameter estimates of meat quality traits, daily gain, longissimus muscle area, backfat thickness, and intramuscular fat content in Duroc pigs selected for average daily gain. J. Anim. Sci. 83, 2058–2065. Vangen, O., 1980. Studies on a two trait selection experiment in pigs. IV. Estimated maintenance requirements from feeding experiments. Acta Agric. Scand. 30, 143–153. Visser, D.P., 2004. Structuring of Breeding Objectives in the Pork Supply Chain in South Africa. University of Pretoria, South Africa (Ph.D. thesis). Von Rohr, P., Hofer, A., Kunzi, N., 1999. Economic values for meat quality traits in pigs. J. Anim. Sci. 77, 2633–2640. Wolfova, M., Wolf, J., Pribyl, J., 2007. Impact of milk pricing system on the economic response to selection on milk components. J. Anim. Breed. Genet. 124, 192–200.