| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
* Dept. of Animal Science, Iowa State University, Ames 50011;
and
National Pork Board, Des Moines, IA 50306;
and
Dept. of Animal Sciences, University of Nebraska, Lincoln 68583;
and
Dept. of Animal Science, North Carolina State University, Raleigh 27695;
and
¶ Dept. of Veterinary Clinical Sciences, Purdue University, West Lafayette, IN 47907;
and
# Dept. of Animal Science, Kansas State University, Manhattan, 66502; and
and
** Abilene, KS 67410
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResults and DiscussionImplicationsAppendixLiterature Cited |
|---|
Key Words: Animal Genetic Resources • Design • Genetic Variation • Standards • Swine
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionImplicationsAppendixLiterature Cited |
|---|
Many of the breed evaluations and crossbreeding experiments that have been reported were conducted 20 to 30 yr ago (Johnson and Omtvedt, 1973
; Johnson, 1981; Wilson and Johnson, 1981). These experiments compared only the largest purebred populations, and few included breeding company lines. Few production traits were evaluated extensively, and meat and eating quality traits were generally not included. Because selection objectives and populations change genetically over time, data from past research may not accurately reflect present-day differences.
Although some data on breeds and lines are available from private production recording systems, these data are not adequate to compare seedstock populations. These programs do not require any sampling of lines and breeds and do not require a uniform test environment to ensure unbiased genetic comparisons. Comparison of these records could easily result in genetic differences that are confounded with environmental effects.
The Terminal Sire Line Genetic Evaluation Program was designed and conducted in the United States by the National Pork Producers Council (NPPC; Des Moines, Iowa), to compare seedstock populations for use in terminal crossbreeding systems. This article outlines details of the program and provides information to researchers on potential application to future evaluations.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionImplicationsAppendixLiterature Cited |
|---|
Producer leadership within the NPPC appointed a Genetic Evaluation Task Force of six producers and six geneticists in 1990. The mission of this task force was to "evaluate the commercial pork producer’s needs for genetic information and design comprehensive evaluation programs to provide sound information to the commercial pork industry" (NPPC, 1991). This task force had many public meetings to gather industry comments and design scientifically sound programs for producer goals.
Four points of consensus guided the task force and succeeding Genetic Programs Committees (GPC): 1) to provide unbiased, clearly presented results of genetic evaluations to producers of all business sizes; 2) to compare seedstock populations for crossbreeding use instead of pure line use; 3) to use industry resources to reduce program costs and increase industry participation; and 4) to reduce environmental differences, particularly related to health, among seedstock sources entering the program.
Seedstock Populations
Seedstock populations that met the definition of a freely interbreeding population of pigs were needed to accomplish the objectives of this experiment. Broadly defined, a seedstock population is a resource population of boars used to test a reference population of commercial sows. A seedstock population can encompass pure breeds and synthetic breeds but must be a distinctly different source of male germplasm that is distinguishable from other populations. Upon repeated sampling, seedstock populations must provide samples of similar genotypes.
The following criteria were set by the GPC to define a seedstock population for the program: Of the litters produced in the last 5 yr, 90% had dams that were produced within the population and 90% had sires that were produced within the population. Of the litters produced in the most current year, 90% had dams that were produced within the population and 90% had sires that were produced within the population.
This defined pure breeds and breeding company synthetic breeds. For the purposes of this program, lines were pure or crossbred/hybrid seedstock populations. The corporate managers of the lines (i.e., breed association or breeding company personnel) entered pure lines into the program. Crossbred lines, such as F1 or F2 animals, were entered by corporate managers of the parent lines. The parent lines of these crossbred lines were required to meet the definition of a seedstock population.
It was the responsibility of the seedstock breeder to document the genetic history of their population if requested by the GPC. Minimum requirements were three-generation pedigrees of all litters born in the last 5 yr.
Seedstock Sampling
The terminal sire line evaluation required a large number of boars to be tested with sufficient progeny to be assured that significant differences would be detected if they exist. Genetic relationships among boars were limited to half sibs or greater to ensure a wide range of variability. Breeders were prevented from any contact with test pigs to avoid bias of test results.
The objective was to test sufficient animals to ensure that if differences between sire lines were detected, true differences did exist. High levels of statistical accuracy for program results were established: the ability to detect differences of 0.25 SD per trait, a 75% power of test, and a 5% significance rate were selected.
Design of Program
Figure 1 outlines the structure of the complete evaluation program.
|
Eight contemporary breeding groups (1 wk in duration) were established, and semen collected at one of eight commercial boar studs was distributed to cooperating commercial producers. Semen from individual boars was distributed across breeding groups and farms to ensure connectivity across farms and breeding groups. Two doses of semen were sent for each sow entered in the program. Pigs farrowed from each contemporary breeding group made up a separate nursery and finishing contemporary group.
Participating commercial producers were required to meet the following standards: 1) the herd must be pseudorabies federal-monitored or validated-free status; 2) there must be no swine dysentery; 3) no clinical signs of Porcine Reproductive and Respiratory Syndrome; 4) no active transmissible gastroenteritis or other contagious disease prior to collection of pigs; 5) single genetic type in sow herd (terminal crossbreeding); 6) ability to use AI and identify litters; 7) willing to provide to the program one or two pigs per litter at feeder pig prices; and 8) sows should be grouped by genetic type for breeding within 1 wk.
Pigs were purchased from producers at 8 to 23 d of age and transported to a segregated early weaning (SEW) nursery. Pig group health status was standardized in the SEW program. Pigs were moved to two environmentally controlled testing stations with partially slotted-floored pens at 20.7 kg of weight. Pigs started the test at 29.5 kg and completed the test at 113.6 kg.
Pig Management
The GPC implemented SEW technology to standardize health among the test pigs originating from many source herds. The goal of the SEW procedures was to minimize the use of medications and to standardize pig health into a single, high-health status and not necessarily to eliminate all diseases. Pilot projects to evaluate SEW procedures were completed prior to implementation of the program (Goodwin et al., 1993).
One to four pigs per litter from each breeding group were delivered to one of four SEW nurseries on a single day to form a contemporary group. Transportation distance from the commercial herds to the SEW nurseries ranged from 80 to 1,000 km. During transport, four pigs were placed in a 0.6- x 0.6- x 0.4-m plastic ventilated box bedded with wood shavings or paper. Boxes were stacked in a covered van for transport. Temperature monitors were used to monitor air temperature surrounding the pigs during transit. Three SEW nurseries had 1.2- x 1.2-m pens, and pigs were allowed 0.15 to 0.19 m2/pig. One nursery had 1.5- x 3.0-m pens and pigs were allowed 0.23 to 0.28 m2/pig.
Table 1 gives the SEW entry protocol that was used. Ambient air temperature in the SEW nurseries was maintained at 29.5 to 31.0°C at pig level for 10 d after entry. After d 10, air temperature was decreased 0.28°C/d until the air temperature reached 22.0°C. Specific health situations were treated by an attending veterinarian.
|
gives two diets that were fed during the nursery phase of the program. Diet 1 was fed until pigs weighed approximately 8.2 kg. Diet 2 was fed until pigs completed the nursery phase at approximately 20.7 kg. Both diets were pelleted and contained Mecadox (Pfizer, Inc., New York, NY) at 55 mg/kg.
|
Feeding of nursery Diet 2 was continued at the testing station for the first 7 d, and pigs were changed to a grower diet on d 8. Pens were placed on test when pigs weighed an average of 29.5 kg. Table 3 gives the diets used in the grow-finish phase. Diet 1 was fed from approximately 22.7 to 68.2 kg, and Diet 2 was fed from 68.2 kg to market weight.
|
gives definitions of the growth, efficiency, and carcass traits that were evaluated. Pigs were weighed and marketed weekly upon reaching an off-test weight of 112 kg. Leg soundness was evaluated prior to slaughter by a trained panel of pork producers. Leg movement and leg structure were each evaluated on a five-point scale and these scores added to give the soundness score. A score of 10 is ideal. Carcass data were obtained at two commercial abattoirs. A three-rib section of loin (10th to 12th ribs) was removed from each carcass and taken to a university meat laboratory for meat quality and eating quality evaluation. Table 5 gives definitions of the meat quality and eating quality traits that were evaluated.
|
|
|
Results and Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionImplicationsAppendixLiterature Cited |
|---|
to 8 was used to determine the number of animals that needed to be tested to meet specification standards for each trait. Size of differences to be detected are given in Table 6 as a percentage of the SD, which allows Table 6 to be used for all traits. Values in Table 6 are based on a single comparison of a pair of means. For comparisons of multiple means, appropriate methods should be used. If more comparisons are made, results will be less conclusive unless larger numbers are tested.
|
|
. Actual differences used in the program, expressed as a percentage of the SD (25%), are 3.25 d for days to 113.6 kg, 0.0227 kg/d for ADG, 0.0625 g/kg for feed efficiency, and 0.0635 cm for backfat depth. The number of observations needed for any level of testing depends on the ratio of differences to be detected and the variation in the trait.
A basic assumption in Table 6
is that only one progeny per sire is tested. If more than one half-sib progeny per sire is tested, the numbers in Table 6 must be augmented by using the values shown in Table 7.
|
). From Table 6, a 20% difference as a percentage of the SD at the 5% significance level and under the 80% power of test means that 413 animals per group are needed. Table 7 is used if more than one half-sib progeny per sire is to be tested. For litter size, testing 10 half-sibs per sire requires 632 animals (1.53 x 413) per sire line group. Genetic population size in this study was determined by the average number of paternal half-sib families in 2 yr, 1991 and 1992, prior to the initiation of the trial. Only sires represented by three or more litters were considered to be half-sib families when determining the number of boars to be sampled. The design of the program was to have each sampled boar sire four litters. One pig from each of these litters would be tested. For large populations, a minimum of 85 boars entered and 340 progeny tested was required.
From Table 6, differences of 0.25 SD per trait, 75% power of test, and a 5% significance rate could be achieved with 242 observations with one progeny per sire. The number of required observations was adjusted using data in Table 7. Average daily gain and backfat require the greatest increase in number of observations for four progeny per sire (140%), resulting in 340 observations (85 sires with four pigs tested per sire). The Yorkshire, Hampshire, and Duroc breeds qualified as large populations.
Table 8 shows the number of paternal half-sib families per year (number of sires with three or more litters per year) and the number of boars with tested progeny needed before results would be published. The minimum population size was set at 46 sires (100 paternal half-sib families in the two most recent years). If a genetic population was too small to provide 46 boars that had genetic relationships of less than half-sibs, more closely genetically related boars could be used with GPC approval. The number of boars to be entered for smaller populations was calculated by the following formula (derivation given in Appendix):
 |
where Ns is the number of sires in a small sample mean, N is the number of boars a large genetic population would enter (85), and Nt is the total number of sires in a small population from the past 2 yr.
Nine of the 11 sire lines originally entered in this program completed the sampling requirements for statistical analysis. Two lines did not submit adequate numbers of boars to meet program requirements.
Table 9 gives the number of boars sampled, litters born, and test pigs per sire line. Pigs farrowed in 136 commercial herds were tested in the program. Twenty producers had multiple sow genetic types in their herds, and the remaining herds had only one genetic type in their herd. There were 45 sow genetic types reported, and these genetic types were grouped into 11 classes for the trait evaluations. Table 10 gives the number and percentage of litters produced by sow genetic type. Sire lines were used randomly across commercial herds, and all matings were made to produce 100% heterosis in the test pigs.
|
|
and 12, respectively. The original sampling goal was to test one pig from each of four litters per sire. The requirement that boars had to be less than 15 mo of age to enter the program resulted in highly variable amounts of semen produced by each boar. Semen from boars was collected and distributed during several breeding periods to increase the probability of litters produced. Low semen volume and quality restricted the doses of semen produced by several of the young boars, resulting in fewer than four litters by 75.4% of sires used in the program. Use of semen from boars that produced large volumes of semen was limited, and progeny tested by boars with lower semen output was maximized. The percentage of dams that produced one, two, three, and four tested progeny were 32.9, 50.9, 14.8, and 1.6%, respectively. The average number of tested progeny per sire was 4.83 and per dam was 1.85.
|
|
lists the pig performance by contemporary group in the SEW nurseries. Average daily gain for the eight contemporary groups over a 40-d test period ranged from 341 to 405 g/d. Mortality rates in the nursery period ranged from 0.93 to 3.08%. Average grow-finish performance by contemporary group is given in Table 14. Average daily gain for the eight contemporary groups ranged from 803 to 900 g/d and days to 113.6 kg ranged from 165 to 180. Mortality rates in the finisher phase ranged from 0.10 to 2.18%.
|
|
Characterization of numerous beef cattle breeds has been accomplished through the Germplasm Evaluation Program as reported by Koch et al. (1976; 1979; 1982) and Wheeler et al. (1996). Breed differences in production traits have been identified as important genetic resources for improving efficiency, composition, and quality. Evaluation of carcass traits and meat palatability has been completed to assist in determining the potential value of these alternative genetic resources (Wheeler et al., 1996). Large differences among and within sire breeds can be exploited by producers to improve carcass and quality traits and increase the rate of genetic improvement.
Random sample testing has been used very successfully by the poultry industry (Anderson, 2001; Hartman, 1985; Working Group 3, 1999). It provided data for unbiased comparisons of performance of commercial poultry stocks. Strict and representative sampling procedures were established as fundamental prerequisites for unbiased comparisons. The first random sample egg-laying test established in California in 1947 was followed by rapid growth in the number of testing stations for egg-laying and broiler and turkey stocks (Hartman, 1985). Some programs are still in existence today (e.g., the North Carolina layer performance program has been in existence since 1957). The European Community continues to publish a combined summary of information from several countries (Working Group 3, 1999). Without question, random sampling had an impact on the rate of improvement in poultry production efficiency since objective information concerning the relative quality of commercial stocks was readily available.
A very high level of statistical accuracy and extensive genetic sampling of lines was built into this program so that producers could have confidence in using the results to change their breeding programs. The results apply to crossbreeding use because that is how commercial producers use seedstock.
|
Implications |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionImplicationsAppendixLiterature Cited |
|---|
|
Appendix |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionImplicationsAppendixLiterature Cited |
|---|
1 and 2. Then CL = 1 - 2 is the contrast of sample means from two large populations. The variance of CL is:
 |
Then V(CL) = V(1) + V(2) because V(µ1) = V(µ2) = 0, and thus covariances of sample means with population means are also zero. It was determined that the number of sires should be 85 (N = 85) to test sample mean differences (1 - 2) with a 5% significance rate and 75% power of test, and samples of two dams per sire and two pigs per dam (n = 4, and total per population = Nn = 340).
All populations are not large, however, so the variance of their mean and covariance of the population mean and sample mean are not zero. Let the population mean and sample mean of a small population be µS and 
S, respectively. The contrast of the mean difference between a large population (e.g., population 1) and the small population is:
 |
The variance of this contrast is:
 |
The variance of any one of the means with Ni sires and n = 4 can be expressed as:
 |
where Ni is number of sires in the sample or in the total population. If nNi is large, the coefficients on the within litter and dam components of variance are small, and the variance of the mean depends mostly on the sire component of variance and the number of sires. As a result, the equation can be reduced to:
 |
Because in a small population, all sires in the sample mean are also in the population mean, there is a covariance between the small population sample mean and the population mean. Let Ns = number of sires in the small sample mean, and Nt = total number of sires in the small population. Then,
 |
A covariance between the means occurs because the Ns sires in the sample are part of the Nt total sires. The covariance between two progeny of the same sire, one in the sample mean and the other in the population mean, is the sire component of variance. With n progeny per sire in each mean, there are n2 covariances within each sire, and Nsn2 total covariances. The divisor for the sample mean is n(Ns), the number of observations in the sample, and the divisor for the population mean is n(Nt). Then, the covariance between small population sample and population means is:
 |
The variance of the contrast of small and large population means is:
 |
Given that N = 85 for large populations, V(CS-L) should equal V(CL). Then,
 |
The next step is to equate the two variances and solve for Ns.
 |
 |
|
Footnotes |
|---|
2 Correspondence: 109 Kildee Hall (phone: 515-294-6728; fax: 515-294-5698; E-mail: tjbaas{at}iastate.edu).
Received for publication March 13, 2002. Accepted for publication June 3, 2002.
|
Literature Cited |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionImplicationsAppendixLiterature Cited |
|---|
This article has been cited by other articles:
|
|
 |
|
  J. P. Cassady, O. W. Robison, R. K. Johnson, J. W. Mabry, L. L. Christian, M. D. Tokach, R. K. Miller, and R. N. Goodwin National Pork Producers Council Maternal Line Genetic Evaluation: A comparison of growth and carcass traits in terminal progeny J Anim Sci, December 1, 2004; 82(12): 3482 - 3485. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
