BREEDING PROGRESS OF NATIONAL SWINE NUCLEUS HERDS

IN TAIWAN

H. L. CHANG, Z. C. KAO, Y. H. HUANG and C. TAI

’ Taiwan Livestock Research Institute, Hsinhua, Tainan,

Taiwan, R.D.C.

“Pig Research Institute in Taiwan, Chunan, Taiwan, R.O.C.

ABSTRACT

Data on 5,325 individually tested boars and 8,346 group tested gilts in south national nucleus herd (SNNH) between 1983 to 1992 were used to measure annually phenotypic and genetic progress of average daily gain (ADG) and feed/gain ratio (FE) from 30kg to 110kg for boars and from 30kg to 90kg for gilts, and backfat thickness (ABF) as well as age at 110kg body weight of boars and at 90kg body weight of gilts.  Based on the phenotypic change of production traits of boars, the superior was  ranked in order of Landrace, Duroc and Yorkshire. When the genetic trend was evaluated,  Duroc boars had more genetic progress on ADG and D110 than the other  breeds  did,  but the genetic progress on ABF was negative in Duroc boars. Although the selection index was based on combined performances of ADG, FE and ABF since 1983, three breeds had no genetic improvement on FE annually in boars. Positive phenotyic improvement was not oberved on ADG of Yorkshire gilts and no genetic progress on ABF of Duroc gilts either.

A total of 3,799 litters farrowed by 1191 sots during 1983-1992 in SNNH were analyzed. Numbers of piglets born, born alive and at 3-week-age per litter in 1992 were: (9.35, 7.98 and 7.71), (9.87, 8.03 and 7.72) and (9.15, 7.49 and 6.62) for Landrace, Yorkshire and Duroc breeds, respectively. The corresponding annually phenotypic changes over the past decade for litter traits were close to zero and it eight be  due  to  no intention  selection for reproductive  performances. The North National Nucleus Herd had better phenotypical progress on litter performance from 1985 to 1990 as compared to that of SNNH.

The elimination of Halothane stress gene by Halothane test and blood typing was carried out at SNNH starting frow September of 1990. Currently, the percentage of piglets with Halothane positive reaction decreased to almost zero in three breeds. In practice, the genetic marker-assisted selection scheme used at SNNH could provide a breeding model to private breeding farms.

I. INTRODUCTION

National Swine Breeding Committee organized by the Council of Agriculture had set up breeding schemes for two national nucleus herds located at northern Chunan (NNNH) and southern Hsinhua(SNNH). The aims of national nucleus herds are to select superior purebred pigs, mainly Landrace, Yorkshire and Duroc breeds, to take performance test and to propagate and extend stocks all over the nation.  Three major purposes are as follows:

1. To breed stocks with better adapting to the climate of Taiwan;  and  to have hogs with more lean meat and higher feed efficiency by meeting the requirement of consumers and producers;

2. To demonstrate the enterprising management system. to general hog farmers; and

3. To promote governmental stocks supplying system in order to help  those small scale farmers at the designated hog raising region to produce 3-way cross hogs. Therefore, the difference of breeding level between small farms  and  enterprised companies  can be balanced out.  (Chyr 1987).

The management systems and breeding schemes carried out in two national nucleaus herds are more or less in similar way. Litters were born in individual farrowing crates. At weaning (around 32-38 days of age), piglets were brought to a post-weaning building where they were housed in pens of around 8-10 animals. The creep diet was provided ad libitum to piglets  from  14 days of age. At 12-week-age, pigs with at least 25 Kg of body weight and 12 apparently functional nipples, adequate leg strength, and a satisfactory body development were chosen for performance test based on two males and two to four females per litter.

Boars were tested for production performance in individual pens  from 30  Kg  to 110 Kg of body weight and gilts were tested in group from 30 Kg to 90 Kg of body weight.  All pigs were fed at libitum during test period. The backfat thickness was probed at the end  of  testing  in  both  sexes. Average  daily gain (ADG) and feed/gain ratio (FE) from 30 Kg to 110 Kg of body weight in boars,  backfat thickness (ABF) and age to 110 Kg and 90 Kg for  boars  and  gilts  were combined onto the selection index at herdmate comparison test.  The selection indexes used before July of 1991 for gilts and boars were as follows:

– For gilts

I = 100 + 180 (ADG -ADG) – 50 (ABF – ABF)

– For boars,

I = 100 + 60 (ADG – ADG) – 40 (FE – FE) – 45 (ABF – ABF)

whereas, after July of 1991,  the index for Landrace and Yorkshire  boars was reset to

I = 100 + 130 (ADO – ADG) – 40 (FE – FE) – 40 (ABF – ABF)

The top 2% and  25% of performance tested boars and gilts, respectively, with satisfactory leg strength were then selected as potential breeding stocks. In order to speed up the breeding progress, the top 5% boars from Central Test Station yearly could be introduced into nucleus herds.

II. THE IMPROVEMENT OF PRODUCTION TRAITS

A total of 2,140 Landrace, 1,424 Yorkshire and 1,761 Duroc boars tested at individual pen;  and 4,321 Landrace,  1,893 Yorkshire and 2,132 Duroc gilts tested at group basis from south National Nucleus Herd  (SNNH) during  the  period of 1983 to 1992 were analyzed to predicted annual phenotypic and genetic improvement of production traits: average daily gain (ADG,  from 30kg to 110kg for boars and from 30kg to 90kg for gilts), backfat thickness (ABF,  adjusted at 110kg  for  boars  and  at 90kg for gilts),  feed/gain ratio (FE) for boars and age at 110kg (9110) for boars as well as age at 90kg (D90) for gilts.  The evoluation of number of boars and gilts tested are represented on Fig.  1 and 6.  Landrace breed is used as dam line in 3-way cross hog production  system,  so the averages of number of pigs tested yearly in both sexes were larger than the other two breeds which were almost even in numbers of tested pigs.

Multivariate mixed linear models techniques (Henderson,1984) were used for analysis.  Restricted maximum likelihood (REML) via the EM algorithm (Dempster et al.,  1977) was used for estimating the necessary variances and covariances for additive  genetic effects and residuals. Best linear unbiased prediction of genetic values and best linear unbiased estimates of marginal means by year of birth were obtained solving mixed-linear equations, conditionally on the REML estimates of variances and covariances.  The statistical model entertained for ADG, FE, D110 and D90 had taken  into account not only fixed effects of year of birth, of testing season,  of parity and of litter size which pig was born with body weights at test  starting and finished as covariates,  but also random effects of animal and residuals.  The model employed for ABF included all effects as above excluding the body weight at test starting.

The variations of the main productive performances in tested boars and gilts by year of birth from 1983 to 1992 are shown in Fig. 2–5 and Fig. 7–9, respectively.  The main results indicated that the  selection criteria  used was  most  beneficial  to Landrace breed in both sexes for traits considered.  Also,  the production performances of Yorkshire breed were  superior to those of Duroc breed in boars.  However,  it was not the case in gilts as shown in Fig. 7 – 9.

The predicted annual phenotypic change of tested boars and gilts from 1983 to 1992 are listed in tables 1 and 2,  respectively..  Over  the  last ten years, the means of phenotypic yearly improvement on ADG, ABF and D110 in Landrace, Yorkshire and Duroc boars annually were (15.79gm, -0.29mm and -1.96day),  (-1.81gm,  -0.10mm and  0.52day)  ard (2.94gm,  -0.15mm and %.37day),  respectively.  Comparisons on phenotypic changes of production traits in boars indicated that Landrace,  Duroc and Yorkshire performed as they were.  However, Duroc boars had more genetic progress on ADG and D110 than the other breeds did.  However, the genetic  progress on ABF in Duroc boars was negative.  There  was  no phenotypic  or genetic  progress on FE of boars.  When daily gain of gilts was considered, the  positive  genetic progress was found in all three breeds. Although similar results was shown in  predicted annual  phenotypic progress in backfat thickness adjusted at 90Kg weight, the yearly genetic improvement was not different from zero in all breeds considered.

The breeding scheme based on selection index described above at  SNNH was practised in NNNH.  The results are listed in tables 3 and 4 for boars and gilts, respectively.  There were 3,771 boars and 8,020 gilts tested at NNNH during 1986-1991 and during 1985-1990, respectively. Landrace boars had more phenotypical progresses on ADG and FE which were consistent with those found at SNNH.  The similar trend was also observed in gilts at NNNH although it was not the case in gilts at SNNH.

III. THE ELIMINATION OF SENSITIVITY TO HALOTHANE GENE IN SNNH

Quality control of pork in Taiwan is important to swine industry. The physics and chemistry of hog caress postmortem can be affected by an autosomal locus (Hal) with two alleles N and n.  Pigs  with  genotypes  NN and  Nn are not sensitive to Halothane test (Hal-) and are non-susceptible to malignant hyperthermia (MH) under various stress conditions, while pigs with genotype nn are sensitive to halothane (Hal+) and are susceptible  to NH.  Despite  the halothane gene’s favorable effect on lean content,  the Hal-n allele results in a net loss in pig production as the Hal+(nn)  pigs suffer  more  death  losses during the fattening period or transportation, and have lower litter size as well as are highly liable to develop PSE  or DFD meat postmortem.

In order to reduce the proportion of halothane sensitive pigs,  the SNNH had a  two-year screen project  for  Halothane gene starting from September of 1990.  After two years with Halothane screen test,  the percentages  of  piglets  with  Hal+(nn) are close to zeros in all breeds. The variation of frequencies for halothane positive piglets by  the  month of  test  within breed is presented on Fig.  10A. However,  the halothane test has difficulty to identify genotype NN and Nn pigs.  Phosphohexose Isomerase  (PHI)  and  6-Phosphogluconate  dehydrogenase  (PGD)  of  blood markers have been used to help identify the  genotypes  after  October  of 1991. The frequencies  of  blood  type (AA, hB and BB) for PHI and PGD within breed are shown in Fig. 108. The  results indicated  that  PHI-A allele was  linked to PGD-B  allele tightly in Landrace breed and PHI-A allele is linked to PGD-A allele in both Yorkshire and Duroc breeds.

IV. THE IMPROVEMENT OF .l.lTTER TRAITS

Sow reproduction  is an economically important aspect of pig production and improving  sow productivity is a mjor way to increase the efficiency in pig production system (Tess et al.,  1983;  Legault,  1985; 1989).  Litter  size  is recognized as a major factor affecting the saws’ productivity in terms of number  of  piglets  weaned per sow per year. However, increasing litter size by genetic or environmental means has been difficult at best (Cunningham et al., 1979; Ollivier, 1982) due to its low heritability and repeatbility.  Under the concept of above,  the traits of interest and of analyzed were:  number of piglets born (LSO),  number  of piglets born alive (LSA) and number of piglets at 3-week-age (LS3).

Data of 3,799 litters of Landrace,  Yorkshire and Duroc breeds produced by 1,191 sots  during  1983-1992 at SNNH were analyzed.  Three litter traits (LSO, LSA and LS3) were described by mixed linear models and analyzed using the methods described in the analysis of production traits. The model employed first included year of farrowing, season of  farrowing and  littermate size effect when the sow was born, as fixed effects, and additive genetic effects as random variable.  Because none of the ratio of additive genetic variance to corresponding residual variance in traits considered was larger than 10^-8 when evaluated at the REML estimates of the components of variance,  the additive genetic effects was eliminated. Therefore, the model used for analysis of litter traits became fixed.

The summary variation of litter  performances  are  represented in Fig. 11  for LSO, LSA and LS3,  and the variations of litter traits within breeds are shown in Fig. 12–14.  In summary,  Duroc breed had one piglet less  at  3-week-age as compared to that of Landrace or  Yorkshire  breed over  the last decade.  The estimated  annual phenotypic change of litter performance at SNNH is listed  in table 5.  During last ten years,  the annually phenotypic changes for litter size at 3-week-age were  positive with the exception  of  Landrace breed.  However,  none of yearly changes in litter traits considered  was significantly different from zero in each breed at SNNH.  The similar result was also found in overall evaluation for traits considered in the same herd.

More than 2,500 litters produced by NNNH were used to evaluate litter performance and corresponding phenotypic improvement.  The statistical model employed was fixed effect model which included year of farrowing, season of mating,  parity and interaction between  season  of  mating  and parity.  The variation of litter traits as mentioned above from 198l to 1992  are shown in Fig. 15.  The results indicated that survival  rate  at 3-week-age was about 82% to 85% with the exception of 1987.  The estimates of  phenotypic improvement yearly for litter traits are listed in Table 6. Landrace breed had more phenotypical progress on litter size and litter weight.

V. IMPLIGATION

National Swine Nucleus Herd at North and at South  in Taiwan have presented an example of breeding scheme  for genetic improvement of production and reproduction traits in Landrace, Yorkshire and Duroc breeds for the last decade by using a combined index of ADG, FE and ABF. Although the genetic improvement for each  trait among breeds  was inconsistent with few exception,  the reproduction performance of sows showed no significant decreases in litter size and weight.  In conclusion, National Nucleus Herds have demonstrated a selection scheme with emphasis on growth performance test for private breeding farms and will recommend the  necessary of setting up a national program on genetic marker-assisted selection for elimination of halothane stress gene.

ACKNOWLEDGEMENTS

We are grateful to Dr.  M.  C.  Wu  for  his  kind  assistance and contribution  of  halothane  test data.  Ifr.  D.  Y.  Lin and Mrs.  H.  R. Tsai are thanked for their skillful help in preparing the manuscript.

REFERENCES BlBLIOGRAPHIQUES

Bidanel,  3.  P.,  J.  C.  Caritez and C.  Legault.  1989.  Estimation of crossbreeding  parameters  between Large  White  and  Eeishan porcine breeds.  I.  Reproductive  performance.  Genet.  Sel.  Evol. 21:507%26.

Chyr, Shuang - Ching.  1987. Pig breeding and development in Taiwan.  The Proceedings of Pig Breeding and Development in Asia.  April 26 – May 3, 1987, ASPAC FFTC, pp.C1-45.

Cunningham,  P.  J.,M.  E.  England,  L.  D.  Young and D.  R.  Zimmerman. 1979.   Selection  for ovulation rate in swine:  correlated response in litter size and weight.  J.  Anim.  Sci.  48: 509-516.

Dempster,  A.  P.,  N.  M.  Laird  and  D.  B.  Rubin.  1977.  Maximum likelihood  from  incomplete  data via the EM algorithm.  3.  R.  Stat. Soc.(B) 39:1-38.

Henderson, C.  R.  1984.  Application of linear models in animal breeding. University of Guelph, Ontario, 462pp.

Legault,  C.  1985.  Selection of breeds, strains and  individual pigs for prolificacy.  J.  Reprod.  Fert.  Suppl.  33:1S1-166.

Ollivier,  L.  1982.  Selection for prolificacy in the pig.  Pig News  and Information, 3:383-388.

Fig.1 The evolution of numbes of boars tested for Landrace, Yorkshire, Duroc breeds in SNNH

Fig. 2 Average daily gain (ADG)by year of birth for boars tested in SNNH

Fig.3 Average backfat thickness at 110kg (ABF) by year of birth for boars tested in SNNH

Fig.4 Average feed/gain ratio by year of birth for boars tested in SNNH

Fig.5 Average age a 110kg (D110) by year of birth for boars tested in SNNH

Fig. 6 The evolution of numbers of gilts tested for Landrace, Yorkshire, Duroc breeds in SNNH

Fig.7 Average daily gain (ADG) by year of birth for gilts tested in SNNH

Fig. 8 Average backfat thickness at 90kg (ABF) by year of birth for gilts tested in SNNH

Fig.9 Average age at 90kg (D90) by year of birth for gilts tested in SNNH

Fig. 10A Frequencies of halothane positive piglets In Landrace, Yorkshire and Duroc breeds during September 1990 to September 1992 in SNNH

Fig 10B. Frequencies of blood type (AA, AB, BB) of PHI and PGD in Landrace, Yorkshire and Duroc pigs finishing performance test during October 1991- September 1992 in SNNH

Fig. 11 Variation of number of piglets born (LS0), born alive (LSA) and at 3-week-age (LS3) per litter in SNNH

Fig. 12 Number of piglets born per litter for Landrace(L), Yorkshire(Y) and Duroc(D) in SNNH

Fig.13 Number of piglets born alive per litter in SNNH

Fig. 14 Number of piglets at 3-week-age per litter in SNNH

Fig. 15 Variation of number of piglets born (LS0), born alive (LSA) and 3-week-age (LS3) per litter during 1983-1992 in SNNH

Table 1. Predicted annual phenotypic and genetic change of growth performance for tested boars born in SNNH during 1983-1992

Table 2. Predicted annual phenotypic and genetic change of growth performance for tested gilts born in SNNH during 1983-1992

Table 3. Predicted annual phenotypic change of growth performance for tested boars born in NNNH during 1986-1991

Table 4. Estimated annual phenotypic change of growth performance for tested gilts born in NNNH during 1985-1990

Table 5. Estimated annual phenotypic change of litter performance for sows in SNNH during 1983-1992

Table 6. Estimated annual phenotypic change of litter performance for sows in NNNH during 1985-1990