Review of Swine Genetics in the U.S.1

Larry D. Young
USDA-ARS, U.S. Meat Animal Research Center
Clay Center, Nebraska


The swine industry continuously faces the challenge of improving the efficiency of converting feed and other input resources into high quality pork and thus, maintain its competitive position in the market place relative to other meat industries and sources of protein, Research in swine genetics plays a central role in meeting that challenge.

Traditionally the swine genetics research has been conducted by land grant universities and the associated State Agricultural Experiment Station and USDA, Agricultural Research Service. Virtually all of these research stations have cooperated to conduct complimentary research starting with the formation of the Regional Swine Breeding Laboratory in 1937. The Regional Swine Breeding Laboratory served as a prototype for all other regional research projects. The concept of Regional Research projects was developed by the directors of the State Agricultural Experiment Stations (SAES) to bring together scientists from SAES, USDA and other institutions and government agencies to work on a problem that is too labor intensive and/or otherwise too costly for a single SAES to undertake. The NCR-l Committee was established in 1970 at the close of the Regional Swine Breeding Laboratory, The NC-l 03 Regional Project replaced the NCR-l Committee in 1971 and was in turn replaced by NC-206 in 1990. NC-206 expired on September 30, 1995, and a new project is being put in place.

During the 1980s, 18 research stations were conducting; quantitative swine breeding research atone time or another. However, the new project being developed includes only eleven research stations. We have lost contributing research programs from Kansas, Illinois, Minnesota, Oregon, Wisconsin, Missouri, and the Beltsville Agricultural Research Center. Of the remaining eleven, two conduct most of their research on the computer or with industry data, four stations have approximately 100 or fewer farrowings a year. and the remaining five programs average approximately 250 farrowings a year. The Regional Project included approximately 11 scientist years in 1981, and the new project includes only 6.25 scientist years. As you can see, the quantitative swine breeding research being conducted at public institutions has decreased greatly in the last decade. The above refers to resources devoted to traditional quantitative swine genetics research and does not include the 4 or 5 scientist years per year involved in developing the swine genetic map during the last five years or so. Stations have been able to contribute to development of the genetic map with relatively small populations of swine. Application of the genetic map to genetic improvement of swine will require larger populations more typical of traditional quantitative genetic projects which may restrict future contributions from these stations. There has been a tremendous increase in swine genetics research conducted by breeding companies during this same time frame. However, for obvious and understandable reasons, their research results are not readily available to the public.


In 1989 there were less than 50 genes mapped in the pig. Since that time, several activities have come together to generate an explosion in the identification and mapping of porcine genes. These activities include the development of the USDA-ARS effort, the Nordic group effort, and the international PiGMaP gene mapping group. In addition, NRSP-8, the Animal Genome Research Program, was initiated by US DA-CSRS. NC-210 "Mapping the Pig Genome" and the work of the UPS. Pig Genome Coordinator have helped develop an extensive mapping database and sharing of genetic materials. The stations contributing to NC-2 10 are Illinois, Iowa, Kansas, Minnesota, Missouri. Wisconsin, Massachusetts, and USDA, ARS, Beltsville Agricultural Research Center. These stations will continue to develop markers to fill in existing gaps in the maps. These combined efforts have led to the development of three linkage maps (Rohrer et at., 1994; Ellegren et at., 1994; Archibald et at., 1995). While the mapping progress has been excellent, progress has also been made in identifying markers for genes with major effects on production traits synch as back fat (Anderson et at., 1994; Yu et at., 1995) and litter size (Rothschild et at., 1994).

These results suggest great opportunities to discover other genes that affect production and reproduction traits in pigs. In 1994 and 1995, three linkage maps of the swine gnome were published and approximately 1,400 genes and markers have been mapped to date. host traits of economic importance in swine are quantitative in nature and determined by an assortment of unidentified genes and environmental factors. Application of the genetic maps in appropriate segregating populations will allow the identification of markers linked to quantitative trait loci (QTL) that control components of growth, muscle quality, reproduction, and health. Markers closely linked to the QTL can be used in combination with phenotypic information through marker-assisted selection to enhance rates of genetic improvement. Already, QTL discoveries for meat quality and litters size are being used in the pork industry. Identification of QTL will yield valuable information regarding the specific genetic control of reproduction and production traits. Based on recent reports of genes with major effects, these discoveries may also provide candidate genes for introgression or transfer between populations.

Alabama, Iowa, Nebraska, North Carolina, and Oklahoma will study F2 resource populations produced by crossing divergent parental lines or breeds. The F2 of divergent parental stock is expected to be segregating for unknown QTL and identifiable DNA markers, allowing use of phenotypic differences between marker genotypes to locate QTL. In addition, selection lines for reproduction have been developed at Ohio. These various populations will be measured at their respective locations for growth and body composition (all stations) and reproduction (Iowa, Nebraska, North Carolina, and Ohio). At some stations, more specific measurements of production and reproduction will be made (e.g., muscle quality, components of litter size) In addition to the traits that arc measured, the participating stations agree to collect tissue samples from their resource families and/or selection lines. DNA derived from the samples will be used directly by the stations and be shared with other stations as needed to confirm significant findings.

Alabama has a line of Landrace that has undergone six generations of selection for increased weight at 200 days of age, and a contemporaneous control. The lines express significantly different body weight at 200 days of age, and average back fat at market weight. An F2 resource population will be produced by crossing these lines, and growth and body composition of approximately 350 F2 animals will be measured. Genotyping facilities are not available at Alabama, thus, tissue samples will be shared with Iowa.

North Carolina will produce an F2 from lines that have undergone 10 generations of divergent selection for plasma testosterone concentration. The high line is approximately 3x greater than the low for both basal concentration and concentration after GnRH challenge, and selection has resulted in correlated responses in growth, back fat, and litter size. Approximately 350 F2 animals will be measured for growth, body composition, and litter traits, as well as plasma testosterone, but genotyping of the population is not possible at North Carolina. Consequently, DNA from this resource family will be sent to Iowa and typed for various micro satellite and candidate gene markers. The markers studied will be primarily in regions previously identified to contain QTL in other populations.

Nebraska has a resource population that is an F2 derived by crossing an unselected control line and a line developed by selecting on a index of ovulation rate and embryo survival After 10 generations of selection, differences between the select and control lines were 6.7 ova shed (~2.5 std. dev.), 3.3 embryos at day 50 of gestation (~1.25 std. dev.), and 3.1 fully formed pigs at term (~1.2 std. dev.). (grandparents from the index line were selected for high index values, and those from the control were selected for low index values. Approximately 450 F2 gilts will be evaluated at 110 days of gestation for components of reproduction and typed for candidate genes and micro satellite markers. DNA Cram this family will also be sent to the laboratory at Iowa, which is specially equipped to type for additional candidate gene markers previously associated with reproduction.

Oklahoma will produce a resource population that is an F2 of lines resulting from divergent selection for post weaning gains. The lines presently differ by approximately three standard deviations for gain and back fat, and four standard deviations for feed intake. Data for growth, feed intake, body composition, and muscle quality will be collected from approximately 700 F2 animals. In preliminary studies, the parental lines are being typed at Oklahoma for several micro satellite markers, and also at Iowa for candidate genes that are the focus of studies there (e.g. PIT-1, aactinin2). Once the F2 is produced, typing for micro satellite markers will continue at Oklahoma, but the project will benefit greatly from typing for candidate genes at Iowa.

Ohio will develop selection lines based on the discovery of candidate gone markers at other stations. Yorkshire and Large White populations will provide the base for select and control lines. Initially, plans are to select for litter size based on typings for the estrogen receptor locus (ESR). ESR genotyping will be completed at Iowa using samples sent from Ohio. Additional marker selection criteria will be incorporated into the study as discoveries occur.

Iowa has an F2 resource population of more than 350 animals produced from original matings of Meishan x Duroc, Meishan x Hampshire, Meishan x Landrace, Minzhu x Landrace, and Minzhu x Hampshire. Growth and carcass data have been collected on all animals, and collection of reproductive data has begun. DNA and phenotypic data can be shared with all stations. These animals will proved: an important resource for identifying QTL and testing markers determined to be significant in populations at other stations.

When a region of chromosome (single marker) is determined to be linked to a QTL, that marker will be bracketed with other closely linked markers to more precisely locate the QTL. Linkage relationships between marker and QTL alleles, and the relative effects of QTL alleles, may differ by population. Therefore, it is extremely important to test significant markers in as many resource families as possible. The several families available in this project represent a wide variety of domestic and imported breeds and generations of controlled selection. The ability to share QTL information for testing in these populations is a vital step toward application of QTL technology,

As significant markers are identified and tested across populations, the potential impact on expected response to selection will be determined. Based on these results, experiments will be designed and initiated at Iowa, Nebraska, Ohio, Oklahoma, and Indiana (using a population presently under selection for reproduction) to measure response from marker-assisted selection. Thus, the regional collaborations described here will provide tiles QTL information necessary for, and facilitate the start of, the first comprehensive experimental test of marker-assisted selection.


A major research effort in the past has focused on the evaluation of genetic conical of sow productivity, uterine capacity, ovulation rate, and reproductive parameters measurable in males. Research under this objective has focused mainly on the genetic regulation of female reproduction with some attention on male reproduction. Selection has been for traits directly affecting female reproduction (sow productivity index, uterine capacity, ovulation rate, embryo survival) or traits measurable in males that may affect male reproduction or be indicative of reproductive potential of female relatives (testicular size or hormone levels). Breeds and selected lines have been compared to evaluate genetic and physiological conical of ovulation rate, uterine capacity, embryo/fetal survival, litter size, and expression of puberty. Breeds, selected lines, and crosses have been used to determine the degree of additive and nonadditive genetic control on reproduction.

Two stations reported positive response to selection for indexes of sow productivity. The indexes were not identical, but included litter size at birth and 21 days as well as weight at 21 days. Two stations have on-going selection experiments focusing on the components of litter size (ovulation rate, embryo/fetal survival, or uterine capacity). One station has one line selected for an index of ovulation rate and number of fetuses at 50 days of gestation and one line selected in two stages for ovulation rate (stage 1) and litter size at birth (stage 2). At the other station, lines are being selected for increased ovulation rate or increased uterine capacity (number of pigs born in unilaterally ovariectomized-hysterectomized gilts).

When selection for reproduction is based only on female performance, response is greatly affected by the low selection intensity that can be obtained. Endocrine relationships between males and females have suggested that selection for traits measured in males may result in improved reproduction of females. Selection for increased testicular size and/or high and low testosterone level has been conducted at two stations. Selection has been success fill in changing the traits under selection. Ill general, correlated responses have been positive for growth, undesirable for fat thickness, and small, but positive, for female reproduction.

Three stations have cooperated to import and evaluate three breeds of pigs from The People's Republic of China (Meishan, Fengjing, and Minzhu). Results document very early puberty, high ovulation rate, and high litter sidle of the Chinese breeds relative to domestic UPS, breeds. However, the reduced growth rate and poorer carcass composition of the Chinese breeds may be large enough to offset the economic advantage in reproduction tinder current market conditions that emphasize lean content. These Chinese breeds have proven to be very useful in mappings the s vine gnome because they are genetically and phenotypically different from our domestic pigs.

Litter size can be increased by selection, but rate of change has been slow because litter size is a physiologically complex trait determined by several components. Research within Regional Project NC-206 and its predecessor NC- 103 led to development of a mathematical model in which ovulation rate and uterine capacity interact to determine prenatal survival and litter size. The model generates values that agree well with experimental results. Simulation with the model has led to predictions that selection on ovulation rate and uterine capacity can substantially increase rate of response over that achieved from direct selection for litter size. It is important to further test whether the mathematical model adequately explains the biology of prenatal survival and litter size, to have reliable estimates of parameters of components of the model, and to develop practical selection procedures to enhance response to selection for litter size.

Selection experiments to investigate the mathematical model of litter size will be conducted cooperatively at Nebraska and MARC. Lines will be developed with different selection criteria to validate the model, to obtain estimates o f genetic parameters, and to compare alternative selection strategies to increase litter size. Populations at each site are similar and selection criteria are complementary.

At Nebraska, two-stage selection for ovulation rate and then litter size will be practiced. The purpose of the two-stage selection is to put selection pressure on uterine capacity by selection for litter size among females with high ovulation rates. There will also be substantial selection pressure on ovulation rate. Three lines will be derived from lines previously developed in an experiment that evaluated nine generations of selection for an index of ovulation rate and prenatal survival to 50 days of gestation. One selected line will be derived from the selected index line. Another select line and a control line will be derived from the control index line. Ovulation rate is measured by counting corpora lutes via laparotomy at second estrus in non- pregnant gilts. The highest ovulating gilts are mated and the second stage of selection is for litter size to emphasize uterine capacity.

In another experiment at Nebraska, selection in the index line was changed from the index to selection for litter size after the tenth generation of index selection. Selection for increased litter size at birth will continue in this line. The contemporary control line will be maintained with random selection. Given the high ovulation rate in the index line relative to the control (20.4 versus 13.8 eggs), selection for litter size is expected to put selection pressure almost entirely on uterine capacity.

At MARC, a selection experiment is being conducted to validate the mathematical model of litter size. There are three lines: a line randomly selected, a line selected for increased ovulation rate, and a line selected for increased uterine capacity. There are two replicates of each line. Ovulation rate is measured by counting the number of corpora lutes via laparoscopy at approximately 40 days of gestation, Uterine capacity is measured as the number of fully formed pigs born to gilts that have undergone unilateral hysterectomy-ovariectomy at approximately 160 days of age.

During the selection phase at MARC, uterine capacity will be measured only in the line selected for uterine capacity, and ovulation rate will be measured only in the line selected for ovulation rate. During the selection phase at Nebraska, ovulation rate will be measured only in the lines undergoing two-stage selection. To validate the mathematical model, ovulation rate, uterine capacity, and litter size will be measured in the lines at each location. To compare lines developed at different stations, a designated line will be used at both Stations as a common control. Based on these results, the two stations will cooperate in line crossing experiments to determine the relative importance of individual and maternal direct and heterotic effects on ovulation rate and uterine capacity as they interact to determine prenatal survival and litter size.

North Carolina is conducting research for litter size using the Nebraska control line. All litters will be standardized at birth, and boars and gilts from the largest litters at birth will be kept for the next generation. Previous research at North Carolina has indicated that gilts raised in large litters tend to have smaller litters than gilts raised in small litters. Thus, the standardization of litter size is intended to remove this environmental effect. Use of the Nebraska control line will allow comparison with lines previously selected for reproductive traits at Nebraska.


A third major area of research has been the evaluation of genetic control of protein and fat accumulation during growth. Three stations have evaluated growth, efficiency of lean and fat deposition, and carcass composition in crosses involving three Chinese breeds (Meishan, Fengjing, and Minzhu) and crosses of UPS. breeds. Results showed Chinese pigs were lighter at birth and grew slower than Duroc pigs, In addition, the Chinese cross pigs were lighter at all ages with the difference between the Chinese and Duroc crosses increasing as age increased. When slaughtered. Chinese pigs had more internal and external fat, smaller longissimus muscle area, and less weight of closely trimmed lean cuts than the Duroc pigs. Growth and carcass traits were similar between the three Chinese crosses. Comparisons among the 1/4 and 1/2 genotypes showed the same general rankings, but the differences were smaller between the Chinese and Duroc crosses. These results show the economic advantages in increased reproductive performance in Chinese crosses are offset, to some extent, by reduced performance in growth rate and carcass composition in the offspring of Chinese cross sows.

Researchers at two stations have selected pigs in which the sole criterion for selection was rapid growth. At one station, selection was conducted where ad libitum feed consumption was allowed, while at the other station selection was either on an ad libitum or restricted feed intake. Both stations maintained unselected control lines on similar diets. In evaluation of these lines, when pigs were allowed ad libitum access to feed, barrows from the ad libitum and restricted select lines grew faster than the pigs from the control line, but carcass characteristics were similar or deteriorated. When given a restricted quantity of feed. barrows from the restricted feed intake select line had less carcass back fat than barrows from the conical line, but the lines did not differ significantly for percentage of lean or percentage of fat in the carcass. differences between the lines for lean tissue growth rate and lean tissue feed conversion were not significant. These results showed no advantage in improved lean tissue efficiency to selection for growth rate under a limited-feeding regime.

Two stations have studied the effect of the halothane (HAL) genotype on lean and fat deposition. A Pietrain line of pi s that is homozygous normal (Near Pietrain) has been established to compare with a Pietrain line that is homozygous for the halothane gene. The Pietrain had 5% more lean in the carcass than the Near Pietrain line. Comparisons of the homozygous normal, heterozygous, and PSS genotypes show the HAL genotype has significantly lower 45 min pH and intramuscular fat in the longissimus muscle, and subjective quality color, marbling, and texture scores. These results suggest advantages of increased lean production in the HAL Pietrain line are countered by poorer meat quality.


Several swine selection experiments were conducted in the 1970s, and the majority involved selection for some components of growth or carcass merit. These experiments documented the efficacy of specific selection criteria and procedures to genetically change growth rate. lean growth rate, and fat growth rate. In the late 1980s. mixed-model, Best Linear Unbiased Prediction procedures were modified to account for the family structures in swine data sets, These procedures are being used in the swine industry to estimate breeding values or Expected Progeny Differences (EPDs) for reproduction, growth, and carcass traits. These Eons are based on the performance of purebred animals, and research has shown they can be used very effectively as selection criteria to improve performance of the purebred progeny. However, most of the commercial slaughter pigs produced in the UPS. are the result of crossbreeding programs. Commercial producers are using within breed Ends to choose among boars within a breed to improve specific provenance traits of crossbred progeny. There are theoretical reasons why Epics based on purebred performance may not be indicative of crossbred progeny performance or may not be applicable to a wide range of production environments. Thus, research is needed to determine the ability of EPDs based on purebred performance to predict the performance of crossbred progeny in different production environments.

The vast majority of the slaughter pigs produced in the UPS. are the result of a wide array of crossbreeding programs and production environments. Scientists in Indiana and Georgia have developed mixed-model, Best Linear Unbiased Prediction (BLUP) procedures which account for the family structures in the swine industry to estimate Expected Progeny Differences (EPDs)based on purebred performance. Many commercial producers are beginning to use Ends to choose among animals within a breed for use in their crossbreeding program. While Ends are known to be effective selection criteria for improving provenance of purebred. it is not known how well they predict performance of progeny under a wide range of environments or performance of crossbred progeny. Within-breed Ends may not be predictive under these conditions because of genotype-environment interactions, heterosis, or other theoretical considerations. Thus, it is important to experimentally evaluate the ability of within-breed EPDs to predict performance in a wide range of breed combinations and production environments

Differences between actual and predicted performance will be measured on purebred and crossbred progeny raised in different environments. EPD will be used as the measure of predicted performance. Selection decisions will be made on Ends calculated from purebred populations. Georgia and Indiana have active research programs in the development and implementation of across-herd genetic evaluation procedures based on multiple-trait animal model BLUP estimates of genetic merit. Selection of boars in national AI studs that meet the criteria for either high or low litter size or Terminal Sire Index will be done on a quarterly basis by workers at the Georgia station. Each station will select the specific boars they will use from the qualified list and then collect the data from the progeny at their station. Collation of data across stations and coordination of data analysis will be done by the North Carolina station. The criteria of success will be if the phenotypic differences seen in the progeny are equal to the differences in the Ends.

Georgia, Indiana, North Carolina, North Carolina A&T, and Virginia will conduct live animal experiments to examine the response to selection based on across-herd Ends for either reproductive or growth traits across environments and breed combinations. These five stations will use the same boars chosen on one of the two criteria, either high or low Epics for litter size, or high versus low Terminal Sire indexes (based on back fat and growth rate). Georgia and North Carolina A&T will focus on litter size only, while Indiana, North Carolina, and Virginia will evaluate both litter size and Terminal Sire Index.

Georgia, Indiana, North Carolina, North Carolina A&T, and Virginia will use boars with a wide range of Ends to determine the usefulness of the Epics based on purebred animals in predicting differences in performance of offspring across differing environments. These stations utilize different breeding and gestation facilities which reflect different production environments. Georgia and Virginia will utilize indoor crated facilities typical of the Southeast and East. North Carolina and North Carolina ART will utilize both indoor and outdoor breeding and gestation environments representative of the Southeast. Indiana will utilize modified open-front facilities common in the Midwest.

Georgia, Indiana, North Carolina, North Carolina ART, and Virginia will also examine the usefulness of EPDs based on purebred animals in predicting differences in performance of offspring of differing breed combinations. Georgia will utilize offspring from purebred Yorkshires and from Hampshire-Landrace crossbreds. Indiana will utilize of offspring from purebred Yorkshires and Landrace and their F1s. North Carolina will utilize offspring from a rotational crossing system using white breeds. North Carolina A&T will utilize offspring from Landrace females. Virginia will utilize offspring from white-cross females.

Collectively, these experiments will collaborate to address three general issues relative to across herd Ends based on purebred animals. First, they will quantify the actual differences in reproduction and growth. Second, differences will be evaluated across differing breed combinations and production environments. Thirdly, these results will be used to improve genetic estimation procedures when data from both crossbred and purebred progeny are utilized.


Andersson, L., C.S. Haley, H. Ellegren, S.A. Knott, M. Johansson, K. Anderson, L. Anderson-Eklund, I. Edfors-Lilja, M. Fredholm, I. Hansson, J. Hakansson, and K. Lundstrom. 1994. Genetic mapping of quantitative trait loci for growth and fatness in pigs. Science 263:1771-1774.

Archibald, A.L., C.S. Haley, J.F. Brown, S. Couperwhite, H.A. McQueen, D. Nicholson, W. Coppieters, A. Van de Weghe, A. Stratil, A.-K. Winters, M. Fredholm, N.J. Larsen, V.H. Nielsen, D Milan, N. Woloszyn, A. Robin, M. Dales, J. Riquet, J. Gellin, J.-C. Caritez, G. Burgaud, L. Ollivier, J.-P. Bidanel, M. Vaiman, C. Renard, H. Geldermann, R Davoli, D. Ruyter, E. J.M. Verstege, M.A.M. Groenen, W. Davies, B. Hoyheim. A. Keiserud, L. Anderson, H. Ellergren, M. Johansson, L. Marklund, J.R. Miller, D.V. Anderson Dear, E. Signer, A J. Jeffrey, C. Moran, P. Le Tissier, Muladno, M.F. Rothschild, C.K. Toggle, D. Vase, J. Helm, H.-C Liu, A. Ragman, T.-P. Yu, R.G. Larson and C.B. Schmitz. 1995. The PiGMaP consortium linkage map of the pig (Sus scrofa). Mammalian Genome 6:157-175.

Ellegren, H, B.P. Chonvdhary, M. Johansson, L. Marklund, M. Fredholm, I. Gustavsson and L Anderson. 1994. A primary linkage map of the porcine gnome reveals a low rate of genetic recombination. Genetics 137:1089-1100.

Rohrer, G.A., L.J. Alexander, J.W Keele, T.P. Smith and C.W. Beat tie. 1994. A micro satellite linkage map of the porcine gnome. Genetics 136:231-245.

Rothschild, M.F., C. Jacobsen, D.A. Vase, C.K. Toggle, T. Short, S. Sasaki. G.R. Eckardt and D.G. McLaren. 1994 A major gene for litter size in pigs. Proceedings of the 5th World Congress on Genetics Applied to Livestock Production 21:225-228.

Yu, T. -P., C.K. Toggle, C.B. Sehmitz and M.F. Rothschild. 1995. Association of PITI polymorphisms with growth and carcass traits in pigs. J. Anim. Sci. 73:1282-1288. ______________________

1 Some of the material presented here is abstracted from the Regional Project Outline for Regional Research Project NC-103 "Genetic Regulation of Pork Production." The agricultural experiment stations of Alabama, Georgia, Illinois, Indiana, Iowa, Minnesota, Missouri, Nebraska, North Carolina, Ohio, Oklahoma, Virginia, plus the North Carolina Agricultural and State University, and USDA, ARS, Roman L. Hruska U.S. Meat Animal Research Center (MARC). Clay Center, NE, are contributing to this project.