Evaluation of the effect and profitability of gene-assisted selection in pig breeding system

Evaluation of the effect and profitability of gene-assisted selection in pig breeding system 1 first-author 2 3 corresp ‡Corresponding Author ‡Corresponding Author †E-mail:chyaosh@mail.sysu.edu.cn †E-mail:chyaosh@mail.sysu.edu.cn chyaosh@mail.sysu.edu.cn College of Animal Science, South China Agricultural University, Guangzhou 510642, China College of Animal Science, South China Agricultural University, Guangzhou 510642, China College of Animal Science and Technology, China Agricultural University, Beijing 10094, China College of Animal Science and Technology, China Agricultural University, Beijing 10094, China College of Life Science, Sun Yat-sen University, Guangzhou 510275, China College of Life Science, Sun Yat-sen University, Guangzhou 510275, China Objective: 2 To evaluate the effect and profitability of using the quantitative trait loci (QTL)-linked direct marker (DR marker) in gene-assisted selection (GAS). Methods: 2 Three populations (100, 200, or 300 sows plus 10 boars within each group) with segregating QTL were simulated stochastically. Five economic traits were investigated, including number of born alive (NBA), average daily gain to 100 kg body weight (ADG), feed conversion ratio (FCR), back fat at 100 kg body weight (BF) and intramuscular fat (IMF). Selection was based on the estimated breeding value (EBV) of each trait. The starting frequencies of the QTL’s favorable allele were 0.1, 0.3 and 0.5, respectively. The economic return was calculated by gene flow method. Results: 2 The selection efficiency was higher than 100% when DR markers were used in GAS for 5 traits. The selection efficiency for NBA was the highest, and the lowest was for ADG whose QTL had the lowest variance. The mixed model applied DR markers and obtained higher extra genetic gain and extra economic returns. We also found that the lower the frequency of the favorable allele of the QTL, the higher the extra return obtained. Conclusion: 2 GAS is an effective selection scheme to increase the genetic gain and the economic returns in pig breeding. Swine Gene-assisted selection (GAS) Selection efficiency Extra returns INTRODUCTION 1 In recent years, more and more quantitative trait loci (QTL) and their closely linked markers have been identified (Georges et al., 1995 1995 ; Blattman et al., 1996 1996 ; Heyen et al., 1999 1999 ; Schrooten et al., 2000 2000 ). The direct marker (DR marker), the functional polymorphism that can be genotyped directly (Dekkers, 2004 2004 ), is used in breeding programs. Many simulation studies have shown that marker-assisted selection (MAS) can improve the genetic merit of livestock (Lande and Thompson, 1990 1990 ; Meuwissen and van Arendonk, 1992 1992 ; Meuwissen and Goddard, 1996 1996 ; Ruane and Colleau, 1996 1996 ; Gomez-Raya and Klemetsdal, 1999 1999 ). Although there are a few studies investing the profitability that the MAS is used in breeding programs in livestock enterprise (Brascamp et al., 1993 1993 ; Spelman and Garrick, 1998 1998 ; Gomez-Raya and Klemetsdal, 1999 1999 ; Hayes and Goddard, 2003 2003 ), no study on the profitability using gene-assisted selection (GAS) in breeding programs has been reported. While Hayes and Goddard ( 2003 2003 ) assessed the profitability of implementing MAS in pig breeding programs, the population and the construction of the breeding system in their study did not fit reality. Moreover, the effective way to improve the genetic gain is to use the DR markers in GAS (Villanueva et al., 2004 2004 ). There has been no paper on the economic profit of applications of GAS schemes in a pig enterprise, either. Cao ( 2003 2003 ) concluded that 5 traits, including number of born alive (NBA), days to 100 kg body weight (Day), feed conversion ratio (FCR), back fat at 100 kg body weight (BF), and intramuscular fat (IMF), should be selected in pig breeding programs. The purpose of the present study was to investigate the difference of the selection efficiency among these 5 traits using DR markers in GAS and profits of GAS used in a pig breeding system. MATERIALS AND METHODS 1 Population model 2 Three discretely closed pig base populations were simulated using stochastic simulation, including 100, 200, and 300 sows plus 10 boars in each group, respectively. The pigs in the base populations ( t =0) had no inbreeding. The 5 traits under selection were genetically controlled by an infinite number of additive loci, each with an infinitesimal effect (polygenes), plus a single bialleles (alleles B and b) QTL. The total genetic value of the i th individual was G i = q i + u i , where q i is the genotypic value due to the QTL and u i is the polygenic effect. The QTL had an additive effect ( a ), defined as half the difference between the two homozygotes, and a dominance effect ( d ) defined as the difference between the heterozygote and the average of the two homozygotes. Thus, genotypic values due to the QTL were a , d , and − a for individuals with genotypes BB, Bb, and bb, respectively (Falconer and Mackay, 1996 1996 ). The additive genetic variance explained by the QTL in the base population was σ q 2 =2 p (1− p ) α 2 , where p is the initial frequency of the QTL’s favorable allele (B) and α is the effect of gene substitution, α = a −(1−2 p ) d (Falconer and Mackay, 1996 1996 ). So the QTL effect ν i is $$\sqrt {{{\sigma _q^2 } \mathord{\left/{\vphantom {{\sigma _q^2 } {2p(1 - p)}}} \right.\kern-\nulldelimiterspace} {2p(1 - p)}}}$$ , 0, $$- \sqrt {{{\sigma _q^2 } \mathord{\left/{\vphantom {{\sigma _q^2 } {2p(1 - p)}}} \right.\kern-\nulldelimiterspace} {2p(1 - p)}}}$$ for QTL of individual i with genotypes BB, Bb, and bb, respectively. The pigs in generation 1 ( t =1) were obtained from randomly mating pigs in base populations. The numbers of selected sows and boars were constant across generations ( t =1~5). At t =0, the polygenic effect for each individual was obtained from a normal distribution with mean zero and variance σ u 2 . Alleles at the QTL were chosen at random with appropriate probabilities (i.e., those given by the initial allele frequencies). The QTL and polygenes were in linkage phase equilibrium. The phenotypic value for individual i ( P i ) was obtained by adding to the total genetic value ( G i ) a normally distributed environmental component with mean zero and variance σ e 2 . In base population P (phenotypic value) was assumed to equal to 1. In subsequent generations, the offspring was obtained by random mating boars and sows without mating among full-sibs and half-sibs. The polygenic effect of the offspring was generated as the average of the polygenic effects of their parents plus a random Mendelian deviation. The latter was sampled from a normal distribution with mean zero and variance ( σ u 2 /2[1−( F s + F d )/2]), where F s and F d are the inbreeding coefficients of the sire and dam, respectively. The QTL alleles were transmitted from parents to offspring in classical Mendelian fashion. A total of 100 replicates were performed for each simulation. Parameters used in this study were summarized in Table 1 1 . The data are heritability and the ratio of QTL variance to genetic variance; P 0 was the initial frequency of the favorable allele of the QTL Parameters Value Population scale Number of breeding sows (boars) 100 (10), 200 (10), 300 (10) Average litter size 8 Traits for selection * * NBA (Short et al., 1997 1997 ) 0.10, 0.24 ADG (Stearns et al., 2005 2005 ) 0.30, 0.11 BF (Sato et al., 2003 2003 ) 0.52, 0.20 IMF (Dragos-Wendrich et al., 2003 2003 ) 0.47, 0.39 FCR (Ovilo et al., 2002 2002 ) 0.23, 0.23 QTL (direct marker) Number of QTL 1 Number of alleles 2 P 0 0.1, 0.3, 0.5 Simulated generation 5 Genetic model and selection procedure 2 BLUP (best linear unbiased prediction) is a common method to estimate the animal breeding value. The pigs selected for breeding were based on the EBV (estimated breeding value) estimated by three genetic models: SBLUP (standard BLUP) model, QBLUP (QTL-BLUP) model, and FBLUP (fixed-BLUP) model. SBLUP model was used when the QTL effect and polygenic effect were combined into the additive effect of the animal in the mixed model, as Y = Xb + Za + e = Xb + Z ( u + v )+ e , where Y , b , a , e , u and v were the vectors of phenotypic value, fixed effect, additive effect, environmental effect, polygenic effect, and QTL effect, respectively. X and Z were the incidence vectors of vectors b and a . $${\mathbf{V}}\left[ {\begin{array}{*{20}c}{\mathbf{a}} \\{\mathbf{e}} \\\end{array} } \right] = \left[ \begin{gathered}{\mathbf{A}}_a \sigma _a^2 {\text{ }}0 \\0{\text{ }}{\mathbf{I}}\sigma _e^2 \\\end{gathered} \right]$$ , where V is the abbreviate for variance, A a , I were the additive genetic relationship and identity matrix, respectively. In the SBLUP model, total EBV in generation t ( EBV t ) was estimated using the total initial genetic additive variance ( σ q 2 + σ u 2 ) and phenotypic values uncorrected for the QTL effect (i.e., P i ). QBLUP model was used when the QTL effect in the model was the random effect, as Y ′= Xb + Zu + e , where Y ′ was vector of the phenotypic value corrected for the QTL effect, and y i ′= y i − v i , where y i and v i were the phenotypic value and QTL effect of individual i , respectively. The other letters in the model denoted the same meanings as in Eq.( 1 1 ). With QBLUP selection, EBV t = EBV u + BV q , where EBV u is the estimate of the polygenic breeding value and BV q is the breeding value due to the QTL. EBV u for individual i is h u 2 ( P i − q i ), where h u 2 is the polygenic heritability and q i is the genotypic value for the QTL. FBLUP model was used when the QTL effect was fixed effect in the model, as Y = Xb + X g g + Zu + e , where Y , b , g , u , and e were the vectors of phenotypic value, fixed effect, QTL genotypic value (obtained from QTL genotype), polygenic effect, and environmental effect, respectively. X , X g , and Z were the incidence vectors of vectors b , g , and u . $${\mathbf{V}}\left[ {\begin{array}{*{20}c}{\mathbf{u}} \\{\mathbf{e}} \\\end{array} } \right] = \left[ \begin{gathered}{\mathbf{A}}_u \sigma _u^2 {\text{ }}0 \\0{\text{ }}{\mathbf{I}}\sigma _e^2 \\\end{gathered} \right]$$ , where A u was the additive genetic relationship due to polygenic genes. With FBLUP selection, EBV t = EBV u + g . Selection efficiency 2 Selection efficiency (SE) was a value to compare the genetic gain of different selection schemes, SE = EBV QBLUP or FBLUP / EBV SBLUP . Economic model of pig enterprise 2 The pig enterprise was composed by nucleus, multiplier, and commercial populations. Gains from increase in genetic merit in the nucleus were realized from sale of slaughter pigs from a nucleus, multiplier tier and commercial tier. A gene flow method (Hill, 1974 1974 ) was used to calculate the extra returns from GAS. Fig. 1 1 gives the flow of animals between tiers and pigs sent to slaughter for a 6-month period in 100-sow nucleus (culled boars and sows have been ignored for simplification). The population construction, the replacement of breeding pigs, and pigs to slaughter in each tier per 6-month period for 100-sow, 200-sow, and 300-sow nucleus were listed in Table 2 2 , respectively. M:F, M, F, C (M), C (F) were the boar to sow ratio, the replacement boars, the replacement gilts, male pigs to slaughter, and female pigs to slaughter, in each 6 months in each tier, respectively 2 Sows in nucleus 5 Nucleus 5 Multiplier 5 Commercial M:F M F C (M) C (F) M:F M F C (M) C (F) M:F M F C (M) C (F) 100 1:10 10 50 342 133 1:50 10 217 2925 1138 1:50 48 7150 32175 32175 200 1:20 10 100 694 266 1:87 10 434 5850 2275 1:50 96 14300 64350 64350 300 1:30 10 150 1047 400 1:130 10 650 8775 3412 1:50 143 21450 96525 96525 The gene flow model requires a vector ( g ) of additional gains from GAS per 6-month period in the nucleus. By considering one generation to be 18 months, g (dimensions 15×1) was derived from a 5×1 vector of additional gains per generation from GAS from the simulations, the genetic gains in each 6-month in the economic model were 1/3 of the genetic gains in each generation in the simulation nucleus (Hayes and Goddard, 2003 2003 ). In the gene flow model, boars and sows both reached puberty at 6 months old. Boars selected for breeding in the nucleus were used for 6 months, and then transferred to the multiplier herd where they were used for another 18 months. Boars not selected for breeding in the nucleus were either slaughtered or used for breeding in the commercial tier, where they were used for 18 months. In the nucleus, half of the piglets produced were from sows in their first parity, 1/3 from sows in their second parity and 1/6 from sows in their third parity. There were eight piglets weaned per mating in the nucleus. Gilts in the nucleus and not selected for use in the nucleus were slaughtered. In the multiplier tier, it was estimated that 30% of piglets were from sows in their first parity, 26% of piglets from sows in their second parity, 24% of piglets from sows in their third parity, and 20% of piglets from sows in their fourth parity. Gilts produced in the multiplier were either used for breeding in the commercial tier or slaughtered. Boars produced in the multiplier were slaughtered. In the commercial tier, it was estimated that 25%, 20%, 17%, 15%, 13% and 10% of piglets were from sows in their 1st, 2nd, 3rd, 4th, 5th, and 6th parities, respectively. The gene flow method required the definition of tier-sex-age classes (Table 3 3 ). No. Tier Sex Age (months) 1 Nucleus Male 0~6 2 Nucleus Male 6~12 3 Nucleus Male 12~18 4 Nucleus Female 0~6 5 Nucleus Female 6~12 6 Nucleus Female 12~18 7 Nucleus Female 18~24 8 Multiplier Male 0~6 9 Multiplier Male 6~12 10 Multiplier Male 12~18 11 Multiplier Male 18~24 12 Multiplier Female 0~6 13 Multiplier Female 6~12 14 Multiplier Female 12~18 15 Multiplier Female 18~24 16 Multiplier Female 24~30 17 Commercial Male 0~6 18 Commercial Male 6~12 19 Commercial Male 12~18 20 Commercial Male 18~24 21 Commercial Female 0~6 22 Commercial Female 6~12 23 Commercial Female 12~18 24 Commercial Female 18~24 25 Commercial Female 24~30 26 Commercial Female 30~36 27 Commercial Female 36~42 Economic returns for GAS 2 Returns from selection in the economic mode were calculated as (Zhang, 2000 2000 ): $$E = \sum {\sum {\Delta G_{ij} n_{ij} } } W^* ,$$ where Δ G ij was the extra genetic gain for GAS to SBLUP selection, n ij was the standardization trait discount value, and W * was the trait’s marginal profit. The marginal profits W * of the 5 traits, NBA, ADG (average daily gain), FCR, LMR (lean-meat rate at 100 kg body weight, indirect selection from BF), and IMF were 8.06, 0.87, 86.32, 10.78 and 23.1 (modified according to the Chinese market), respectively (Cao, 2003 2003 ). Standardization trait discount value was calculated as (Zhang, 2000 2000 ): $${\mathbf{n}}_{ij} = \sum\limits_{t = 1}^D {h'_i m_{jt} (1 + r)^{ - t} } ,$$ where h i was the realization vector of the trait i , m jt was the vector of gene proportion at the time t in the j selection group in every tier-sex-age group. r was the discount rate and was assumed to be 0.05 here. There were 15 selection groups, so D =15. m jt = Ra t −1 + Pm it −1 , where P (27×27) was the gene transmission matrix. a t = Qa t −1 ( t =1,2,…,15), where a 1 was the initial vector for all selection group, and a 1 ′=[100100010001000010001000000]. t was the selection generation. Q (27×27) was the mature matrix, where the elements of the reproduction rows were 0 and the others were the same as the elements in matrix P . R (27×27) was the reproduction matrix, where the elements of the reproduction rows were the same as those in the P matrix, and the other elements were 0. According to Eq.( 6 6 ), the vector m jt was obtained. Extra returns for GAS 2 The extra cost of the breeding program from implementing GAS was assumed to be only genotyping costs. The genotyping cost for one QTL was RMB 5 yuan. The genotyping costs for three nucleus were C t =8× n ×5× d t , where n was the number of sows in the nucleus, d was the discounted rate, and t was the time in the selection group. So the extra returns of the breeding program from implementing GAS were calculated as: $$R_t = E_t - \sum {_{t = 1}^{15} } C_t$$ . RESULTS 1 Selection efficiency 2 The selection efficiencies (SEs) for the 5 traits using QBLUP model and FBLUP model were presented in Table 4 4 and Table 5 5 , respectively. All SEs of 5 traits at generation 1 were 100%, because in this stage there was no selection for all pigs. Almost all the SEs of the 5 traits in the later generations were higher than 100%, and became lower and lower when the generations went higher. The higher initial frequency of the favorable allele of the selected QTL ( P 0 ) was, the lower the SE. The SE was the highest for trait NBA and the lowest for trait BF. G : Generation; P 0 : The initial frequency of the QTL’s favorable allele 3 G 3 P 0 15 SE 5 100-sow 5 200-sow 5 300-sow NBA FCR ADG BF IMF NBA FCR ADG BF IMF NBA FCR ADG BF IMF 1 0.1 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 0.3 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 0.5 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 2 0.1 1309.3 145.4 114.4 113.0 144.0 26748.0 141.4 109.0 114.0 137.8 500.4 137.1 107.3 111.4 136.0 0.3 155.4 114.2 105.9 106.7 114.7 139.8 112.5 106.3 106.3 115.0 145.8 111.7 103.8 106.7 112.9 0.5 128.3 110.3 104.5 104.5 109.2 118.4 108.4 105.3 104.5 109.3 122.5 108.4 104.1 104.4 108.4 3 0.1 213.2 131.3 114.3 113.0 120.7 233.9 130.7 110.5 113.5 117.4 193.0 124.9 110.4 112.1 115.3 0.3 146.7 108.0 105.5 103.5 109.3 130.1 109.8 105.7 104.7 109.7 133.1 109.9 104.6 105.1 107.5 0.5 126.8 105.5 104.4 104.0 103.5 109.5 106.1 105.0 102.5 105.4 114.4 106.3 103.9 104.6 105.0 4 0.1 151.3 116.7 110.9 107.0 104.2 154.4 115.0 110.6 107.0 103.9 144.5 111.4 109.1 106.5 103.3 0.3 128.4 103.7 104.8 100.9 102.4 114.0 104.3 104.3 102.1 104.1 119.4 104.2 101.9 103.8 102.3 0.5 116.8 101.8 104.6 103.1 101.6 105.7 104.3 103.5 100.5 102.5 108.6 104.0 103.2 102.7 102.0 5 0.1 123.4 106.7 106.7 101.7 100.2 123.6 105.2 107.0 102.4 100.4 125.7 103.6 105.7 102.3 100.4 0.3 120.9 101.0 104.3 100.8 100.5 107.2 101.6 102.4 101.2 101.9 112.9 101.3 100.2 102.5 101.6 0.5 110.7 99.7 103.1 101.6 100.9 100.8 101.6 101.7 99.9 102.8 104.3 102.5 102.0 102.1 101.8 G : Generation; P 0 : The initial frequency of the QTL’s favorable allele 3 G 3 P 0 15 SE 5 100-sow 5 200-sow 5 300-sow NBA FCR ADG BF IMF NBA FCR ADG BF IMF NBA FCR ADG BF IMF 1 0.1 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 0.3 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 0.5 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 2 0.1 1236.1 142.7 112.4 113.5 144.9 23991.8 137.9 106.6 114.1 136.2 465.1 135.3 105.8 111.3 133.9 0.3 153.2 113.6 105.5 106.7 114.7 138.8 110.4 104.3 106.2 114.6 149.8 112.1 103.5 106.4 112.8 0.5 124.9 109.5 104.5 104.7 109.1 120.0 108.5 104.9 104.6 109.4 124.3 108.4 103.7 104.8 108.3 3 0.1 210.2 133.4 113.0 114.7 120.2 230.1 132.0 111.9 114.7 116.2 199.6 124.2 111.5 111.6 114.9 0.3 147.2 111.0 104.4 105.1 109.0 131.4 107.7 104.4 104.5 110.2 135.3 111.6 103.9 105.4 108.1 0.5 124.6 106.2 103.7 102.9 103.4 114.6 106.0 104.2 103.3 105.8 113.9 107.0 103.1 104.6 105.0 4 0.1 157.4 120.1 110.1 107.2 104.2 157.0 116.3 110.8 108.0 102.6 151.0 111.7 109.8 106.1 103.8 0.3 127.3 106.3 103.9 102.3 103.1 115.4 102.1 103.7 102.8 105.5 119.6 105.6 101.5 104.6 102.8 0.5 114.6 102.9 103.4 103.1 102.0 110.0 104.3 103.2 101.5 103.4 106.3 104.1 101.9 102.4 102.6 5 0.1 131.1 108.1 105.2 101.9 99.3 125.0 105.5 107.0 102.4 99.1 128.4 103.3 106.0 101.1 100.8 0.3 118.8 104.7 103.5 101.2 100.8 107.2 99.3 102.6 101.5 103.2 111.9 101.7 100.7 102.9 101.6 0.5 107.6 100.6 101.8 101.9 101.7 103.9 102.1 102.1 100.8 103.8 101.5 102.7 100.7 101.3 101.7 The SE for the sex-limited trait NBA was higher while using QBLUP model in 100-sow and 200-sow nucleus, respectively, but it was higher in 300-sow nucleus when FBLUP model was used. The SE for trait FCR was higher when FBLUP model was used. Most of the SEs for trait ADG were higher when QBLUP model was used. All the SEs for trait BF were higher when FBLUP model was used. The SE for trait IMF was higher when P 0 =0.1 and using QBLUP model in three populations, whereas it was higher when P 0 was 0.3 and 0.5 and FBLUP model was used in three populations, respectively. Extra returns for GAS 2 The extra returns for the three populations at different initial frequencies of the QTL’s favorable allele when two models were used were described in Table 6 6 . The lower the P 0 was, the higher the extra returns were, using GAS in the breeding program. Moreover, the bigger the nucleus population was, the higher the extra returns were, while using GAS. The extra returns fell quickly with the rise of P 0 . P 0 : The initial frequency of the QTL’s favorable allele 3 Model 9 Extra economic returns (RMB yuan) 3 100-sow 3 200-sow 3 300-sow P 0 =0.1 P 0 =0.3 P 0 =0.5 P 0 =0.1 P 0 =0.3 P 0 =0.5 P 0 =0.1 P 0 =0.3 P 0 =0.5 QBLUP 3230647.30 559532.72 86018.29 5599108.49 2243969.43 2231599.75 6726772.58 3453732.47 4955329.35 FBLUP 3768478.93 2414493.20 614107.34 5354031.78 476440.35 3306397.27 6367686.56 4132929.78 5012847.05 The extra returns were higher in 100-sow population when FBLUP model was used, while higher in 200-sow population when QBLUP model was used, excluded when P 0 =0.5. In 300-sow nucleus, the extra returns were higher when P 0 =0.1 using QBLUP model, but lower when P 0 =0.3 and 0.5, respectively. DISCUSSION 1 Factors influencing the selection efficiency 2 The SEs for the 5 traits in the GAS scheme in pig breeding were influenced by the trait’s heritability, characters (such as sex-limited), the ratio of the QTL variance to additive genetic variance, and the initial frequency of the QTL’s favorable allele. A few studies have reported that the SE in MAS or GAS was higher for lower heritability trait and sex-limited trait (Lande and Thompson, 1990 1990 ; Gimelfarb and Lande, 1995 1995 ; Whittaker et al., 1995 1995 ; Ruane and Colleau, 1995 1995 ; Meuwissen and Goddard, 1996 1996 ; Liu et al., 2001 2001 ). Ruane and Colleau ( 1995 1995 ) found that the selected QTL’s effect was a major factor to affect the selection effect of MAS, the lager the QTL effect and the higher the heritability of the trait, the more obvious the selection effect in MAS or GAS. Our result in the present study agrees with this conclusion. Among the 5 traits, NBA, a sex-limited trait, has the highest SE. IMF has the high heritability and QTL variance and its SE was next to the SE of NBA, when P 0 was lower. FCR has both the moderate heritability and QTL variance and its SE was moderate among the 5 traits. BF has the highest heritability and moderate QTL variance and at the first 3 selection generations, its SE was higher than that of trait ADG with the lowest QTL variance. The initial frequency of the favorable allele of the selected QTL ( P 0 ) was a major factor to influence the SE (Spelman and Garrick, 1998 1998 ). The SEs for the 5 traits became lower when P 0 went up. When P 0 goes up the frequency of the QTL’s favorable allele increases more slowly, so the QTL effect and the total genetic effect of QBLUP selection and FBLUP selection increase slower compared with those when P 0 was lower, but faster compared with those of SBLUP selection, which results in a low SE. Such results are consistent with the findings of Liu et al.( 2001 2001 ). The GAS selection effect was negatively associated with the selection generation, with a reduced SE for the higher generation (Hospital et al., 1997 1997 ; Liu et al., 2001 2001 ). With the increase of the selection generation, the frequency of the favorable allele of the QTL increased fast and after 4 generations it almost went to purify, and the QTL variance became lower and lower, then the total genetic gain in MAS or GAS became smaller and smaller. In this study, the SEs for the 5 traits were the highest at the second generation, and decreased with the increasing selection generation. At the fifth generation, although most of the SEs were above 100%, but the effect was small, indicating that for GAS or MAS, 3 to 4 generations were optimal. This study also shows that in the selection of QTL, maximization of response in a short term can result in lower cumulative responses in a long run, which can be solved by using the optimal model (Dekkers and van Arendonk, 1998 1998 ; Dekkers and Chakraborty, 2001 2001 ; Chakraborty et al., 2002 2002 ; Villanueva et al., 2002 2002 ; 2004 2004 ; 2005 2005 ; Tang, 2006 2006 ). The SEs for the same trait had no big difference between two models used, but for the sex-limited trait NBA and trait ADG with the lowest QTL variance, the SEs were higher in QBLUP model. For trait FCR with moderate heritability and QTL variance and trait BF with high heritability and moderate QTL variance, the SEs for both traits were higher when FBLUP model was used. However, for trait IMF, the SEs were almost the same in these two models. Extra economic returns for GAS 2 The extra economic returns for GAS schemes reduced dramatically when P 0 increased, which was consistent with the pattern that the lower the P 0 was, the higher the genetic gain of the selected trait was, using GAS scheme. Among the 5 traits in the pig breeding plan, the trait FCR had the highest margin profit and trait IMF was next to it. According to the gene flow methods, the economic returns direct proportion to the trait’s margin profit (Zhang, 2000 2000 ). So the genetic gains of trait FCR and trait IMF affected the economic returns of the population markedly. So in the pig breeding program, it is important for us to identify the QTL that affect the traits with high margin profit and then clone them. Thus, the economic profit of using GAS scheme would be considerable. Among the three populations, the 300-sow nucleus population had the highest extra profit, and the 100-sow population had the lowest one. Because with the increase of the nucleus size, the total extra genetic gain for the population increased, and the pigs sent to slaughter per 6-month period increased too. In the present study, we didn’t consider that with the increase of the traits’ genetic gain, the production performance increased, so the pigs to slaughter in each six-month period would increase and the extra profit would increase too. With the improvement of the technology to genotype the QTL, the cost of GAS and MAS would decrease, so the extra profit would increase. References AN Blattman , BW Kirkpatrick , KE Gregory . A search for quantitative trait loci for ovulation rate in cattle . Anim Genet , 1996 . 27 157 - 162 . . A search for quantitative trait loci for ovulation rate in cattle EW Brascamp , JAM van Arendonk , AF Groen . Economic appraisal of the utilization of genetic markers in dairy cattle breeding . J Dairy Sci , 1993 . 76 1204 - 1213 . . Economic appraisal of the utilization of genetic markers in dairy cattle breeding HZ Cao . The Breeding Plan for High Quality Pigs , 2003 . China Agricultural University , (in Chinese, Ph.D. Thesis) . The Breeding Plan for High Quality Pigs RL Chakraborty , RL Moreau , JCM Dekkers . A general method to optimize selection on multiple identified quantitative trait loci . Genet Sel Evol , 2002 . 34 ( 2 ): 145 - 170 . DOI: 10.1051/gse:2002001 http://doi.org/10.1051/gse:2002001 . A general method to optimize selection on multiple identified quantitative trait loci JCM Dekkers . Commercial application of marker- and gene-assisted selection in livestock: strategies and lessons . J Anim Sci , 2004 . 82 ( Suppl. ): 313 - 328 . . Commercial application of marker- and gene-assisted selection in livestock: strategies and lessons JCM Dekkers , JAM van Arendonk . Optimizing selection for quantitative traits with information for an identified locus in outbred population . Genet Res (Camb) , 1998 . 71 ( 3 ): 257 - 275 . DOI: 10.1017/S0016672398003267 http://doi.org/10.1017/S0016672398003267 . Optimizing selection for quantitative traits with information for an identified locus in outbred population JCM Dekkers , RL Chakraborty . Potential gain from optimizing multi-generation selection on an identified quantitative trait locus . J Anim Sci , 2001 . 79 2975 - 2990 . . Potential gain from optimizing multi-generation selection on an identified quantitative trait locus M Dragos-Wendrich , G Moser , H Bartenschlager , G Reiner , H Geldermann . Linkage and QTL mapping for Sus scrofa chromosome 11 . Journal of Animal Breeding and Genetics , 2003 . 120 ( s1 ): 89 - 94 . DOI: 10.1046/j.0931-2668.2003.00428.x http://doi.org/10.1046/j.0931-2668.2003.00428.x . Linkage and QTL mapping for Sus scrofa chromosome 11 DS Falconer , TFC Mackay . Introduction to Quantitative Genetics , 1996 . 4th Ed Longman Group Limited Corporation , . Introduction to Quantitative Genetics M Georges , D Nielsen , M Mackinnon , A Mishra , R Okimoto , A Pasquino , L Sargeant , A Sorensen , M Steele , X Zhao , 等 . Mapping quantitative trait loci controlling milk production in dairy cattle by exploiting progeny testing . Genetics , 1995 . 139 907 - 920 . . Mapping quantitative trait loci controlling milk production in dairy cattle by exploiting progeny testing A Gimelfarb , R Lande . Marker-assisted selection and marker-QTL associations in hybrid populations . Theor Appl Genet , 1995 . 91 ( 3 ): 522 - 528 . DOI: 10.1007/BF00222983 http://doi.org/10.1007/BF00222983 . Marker-assisted selection and marker-QTL associations in hybrid populations L Gomez-Raya , G Klemetsdal . Two-stage strategies utilizing marker-quantitative trait locus information and individual performance . J Anim Sci , 1999 . 77 ( 8 ): 2008 - 2018 . . Two-stage strategies utilizing marker-quantitative trait locus information and individual performance B Hayes , ME Goddard . Evaluation of marker assisted selection in pig enterprises . Livestock Production Science , 2003 . 81 ( 2-3 ): 197 - 211 . DOI: 10.1016/S0301-6226(02)00257-9 http://doi.org/10.1016/S0301-6226(02)00257-9 . Evaluation of marker assisted selection in pig enterprises DW Heyen , JI Weller , M Ron , M Band , JE Beever , E Feldmesser , Y Da , GR Wiggans , PM Vanraden , HA Lewin . A genome scan for QTL influencing milk production and health traits in dairy cattle . Physiol Genomics , 1999 . 1 165 - 175 . . A genome scan for QTL influencing milk production and health traits in dairy cattle JM Hill . Prediction and evaluation of response to selection with overlapping generations . Anim Prod , 1974 . 18 117 - 139 . . Prediction and evaluation of response to selection with overlapping generations F Hospital , L Moreau , F Lacoudre , A Charcosset , A Gallais . More on the efficiency of marker assisted selection . Theor Appl Genet , 1997 . 95 ( 8 ): 1181 - 1189 . DOI: 10.1007/s001220050679 http://doi.org/10.1007/s001220050679 . More on the efficiency of marker assisted selection R Lande , R Thompson . Efficiency of marker assisted selection in the improvement of quantitative traits . Genetics , 1990 . 124 743 - 756 . . Efficiency of marker assisted selection in the improvement of quantitative traits HY Liu , Y Zhang , Q Zhang . Relative efficiency of marker assisted selection when marker and QTL incompletely Linked . Chinese Science Bulletin , 2001 . 46 ( 240 ): 2058 - 2063 . . Relative efficiency of marker assisted selection when marker and QTL incompletely Linked THE Meuwissen , JAM van Arendonk . Potentional improvement in rate of genetic gain from marker-assisted selection in dairy cattle breeding schemes . J Dairy Sci , 1992 . 6 ( 75 ): 1652 - 1659 . . Potentional improvement in rate of genetic gain from marker-assisted selection in dairy cattle breeding schemes THE Meuwissen , ME Goddard . The use of marker haplotypes in animal breeding schemes . Genet Sel Evol , 1996 . 28 ( 2 ): 161 - 176 . DOI: 10.1051/gse:19960203 http://doi.org/10.1051/gse:19960203 . The use of marker haplotypes in animal breeding schemes CA Ovilo , JL Clop , MA Noguera , MA Oliver , C Barragán , C Rodríguez , L Silió , MA Toro , A Coll , JM Folch , 等 . Quantitative trait locus mapping for meat quality traits in an Iberianx Landrace F 2 pig population . J Anim Sci , 2002 . 80 2801 - 2808 . . Quantitative trait locus mapping for meat quality traits in an Iberianx Landrace F2 pig population J Ruane , J Colleau . Marker assisted selection for genetic improvement of animal population when a single QTL is marked . Genet Res Camb , 1995 . 66 71 - 83 . . Marker assisted selection for genetic improvement of animal population when a single QTL is marked J Ruane , J Colleau . Marker-assisted selection for a sex-limited character in a nucleus breeding population . J Dairy Sci , 1996 . 79 ( 9 ): 1666 - 1678 . . Marker-assisted selection for a sex-limited character in a nucleus breeding population S Sato , Y Oyamada , K Atsuj , T Nade , S Sato , E Kobayashi , T Mitsuhashi , K Nirasawa , A Komatsuda , Y Saito , 等 . Quantitative trait loci analysis for growth and carcass traits in a Meishan×Duroc F 2 resource population . J Anim Sci , 2003 . 81 ( 12 ): 29 - 38 . . Quantitative trait loci analysis for growth and carcass traits in a Meishan×Duroc F2 resource population C Schrooten , H Bovenhuis , W Coppieters , JAM van Arendonk . Whole genome scan to detect quantitative trait loci for conformation and functional traits in dairy cattle . J Dairy Sci , 2000 . 83 795 - 806 . . Whole genome scan to detect quantitative trait loci for conformation and functional traits in dairy cattle TH Short , MF Rothschild , OI Southwood , DG Mclaren , A de Vries , H van der Steen , GR Eckardt , CK Tuggle , J Helem , DA Vaske , 等 . Effect of the estrogen receptor locus on reproduction and production traits in four commercial pig lines . J Anim Sci , 1997 . 75 3138 - 3142 . . Effect of the estrogen receptor locus on reproduction and production traits in four commercial pig lines R Spelman , DJ Garrick . Genetic and economic responses for within-family marker-assisted selection in dairy cattle breeding schemes . J Dairy Sci , 1998 . 81 2942 - 2950 . . Genetic and economic responses for within-family marker-assisted selection in dairy cattle breeding schemes TM Stearns , JE Beevers , BR Southey , M Ellis , FK McKeith , SL Rodriguez-Zas . Evaluation of approaches to detect quantitative trait loci for growth, carcass, and meat quality on swine chromosomes 2, 6, 13, and 18. I. Univariate outbred F 2 and sib-pair analyses . J Anim Sci , 2005 . 83 1481 - 1493 . . Evaluation of approaches to detect quantitative trait loci for growth, carcass, and meat quality on swine chromosomes 2, 6, 13, and 18. I. Univariate outbred F2 and sib-pair analyses GQ Tang . Study on Optimization Using QTL Information in Selection Schemes , 2006 . Sichuan Agricultural University , (in Chinese, Ph.D. Thesis) . Study on Optimization Using QTL Information in Selection Schemes B Villanueva , BR Pong-Wong , JA Woolliams . Marker assisted selection with optimized contribution of the candidates to selection . Genet Sel Evol , 2002 . 34 ( 6 ): 679 - 703 . DOI: 10.1051/gse:2002031 http://doi.org/10.1051/gse:2002031 . Marker assisted selection with optimized contribution of the candidates to selection B Villanueva , JCM Dekkers , JA Woolliams , P Settar . Maximizing genetic gain over multiple generations with quantitative trait locus selection and control of inbreeding . J Anim Sci , 2004 . 82 1305 - 1314 . . Maximizing genetic gain over multiple generations with quantitative trait locus selection and control of inbreeding B Villanueva , R Pong-Wong , J Fernández , MA Toro . Benefits from marker-assisted selection under an additive polygenic genetic model . J Anim Sci , 2005 . 83 1747 - 1752 . . Benefits from marker-assisted selection under an additive polygenic genetic model JC Whittaker , RN Curnow , CS Haley , R Thompson . Using marker-maps in marker-assisted selection . Genet Res Camb , 1995 . 66 255 . Using marker-maps in marker-assisted selection Y Zhang . Livestock Breeding Plan , 2000 . : Beijing, China China Agricultural University Publishing Company , (in Chinese) . Livestock Breeding Plan I need to thank Dr. Ji-rong Long, Vanderbilt University Medical Center, USA, for her advice and help. 10.1631/jzus.2007.B0822 Project (No. 30300249) supported by the Natural Science Foundation of Guangdong Province, China 2007-09-16 2007-10-08 2022-01-05 Copyright © 2007, Journal of Zhejiang University Science 2007 Journal of Zhejiang University-SCIENCE B (Biomedicine and Biotechnology) 11 08