Effects of errors in pedigree on three methods of estimating breeding value for litter size, backfat and average daily gain in swine

Estimated breeding value (EBV) was calculated based on either individual phenotype (SP), an index of individual phenotype and full- and half-sib family averages (SI) or Best Linear Unbiased Prediction (BLUP). Calculations were done with correct data or data with 5, 10, 15 or 20% of the records per generation containing pedigree errors. Traits considered were litter size (LS), backfat (BF) and average daily gain (ADG). When data were correct, BLUP resulted in an advantage in expected genetic gain over SP of 22, 7.2 or 30.8% for LS, BF and ADG, respectively, and over SI of 9.6, 3.8 or 21.4%. When sire and dam pedigrees were incorrect for 20% of the pigs each generation, genetic gain using SI was reduced by 7, 2.5 or 6.5% and genetic gain using BLUP was reduced by 9.3, 3.2 or 12.4% for LS, BF and ADG, respectively. With 20% of the pedigrees in error, the advantages in genetic gain of using BLUP over SP, the method unaffected by errors in pedigree, were 10.5, 3.8 and 14.6% for LS, BF and ADG, respectively. These results suggest that, although BLUP is affected to a greater degree by pedigree errors than SP or SI, selection of swine using BLUP still would improve response to selection over the use of SP or SI.     

Genetic parameter estimates of meat quality traits in Duroc pigs selected for average daily gain, longissimus muscle area, backfat thickness, and intramuscular fat content

Using a multitrait animal model BLUP, selection was conducted over seven generations for growth rate (ADG), real-time ultrasound LM area (LMA), backfat thickness (BF), and intramuscular fat content (IMF) to develop a new line of purebred Duroc pigs with enhanced meat production and meat quality. This selection experiment examined 543 slaughtered pigs (394 barrows and 153 gilts) from the first to the seventh generation for meat quality traits. Further, electric impedance and collagen content of loin meat were measured from the fourth to sixth generation. The present study was intended to estimate genetic parameters of the correlated traits of tenderness (TEND), meat color (pork color standard: PCS; lightness = L*), drip loss (DL), cooking loss (CL), pH (PH), electric impedance (IMP), and collagen (COL) of the LM, and the genetic trends of these traits. Respective heritability estimates for IMF, TEND, DL, CL, PCS, L*, PH, IMP, and COL were 0.39, 0.45, 0.14, 0.09, 0.18, 0.16, 0.07, 0.22, and 0.23. Genetic correlations of IMF with ADG and BF were low and positive, but low and negative with LMA. Tenderness was correlated negatively with ADG (-0.44) and BF (-0.59), but positively correlated with LMA (0.32). The genetic correlation between LMA and DL was positive and high (0.64). The genetic correlations of TEND with IMF and COL were low (-0.09 and 0.26, respectively), but a moderate genetic correlation (0.43) between COL and IMF was estimated, suggesting related increases of IMF and connective tissue. Genetic correlations among meat quality traits suggested that when IMF increases, the water holding capacity improves. Genetic trends of meat quality traits showed increased IMF and lighter meat color.     

Effects of social interactions on empirical responses to selection for average daily gain of boars

Effects of social interactions on responses to selection for ADG were examined with records of 9,720 boars from dam lines (1 and 2) and sire lines (3 and 4) provided by Pig Improvement Company. Each line was analyzed separately. Pens contained 15 boars. Average daily gains were measured from about 71 to 161 d of age and BW from 31 to 120 kg. Models included fixed effects of contemporary groups and initial test age as a covariate and random direct genetic (a), social genetic (c), social environmental (ce), and litter (lt) effects. Estimates of direct heritability with model 1 (the full model with a, c, ce, and lt) were 0.21, 0.28, 0.13, and 0.15 for lines 1 to 4. Estimates of heritability of social effects were near zero. Estimates of total heritable variance were 55, 52, 38, and 96% of phenotypic variance for lines 1 through 4. Empirical responses to selection with model 1 were calculated using the parameter estimates from model 1. For response of 1 genetic SD for both components (a and c), the proportions of expected total gain due to social effects (with economic weights of 1 and pen size-1 = 14) were 54, 28, 65, and 65% for the 4 lines. Genetic superiorities of the top 10% of boars were calculated for boars ranked using reduced models, but with EBV calculated using the full model (model 1). Average total breeding values (ETBV = EBV(a)+14EBV(c)) for the top 10% of boars selected with model 1 were 74.08, 94.26, 31.79, and 92.88 g for lines 1 through 4, respectively. For rankings based on model 2 (a, ce, and lt), but EBV calculated with model 1, average total breeding values for the top 10% were 68.15, 94.03, 7.33, and 84.72 g with empirical correlated responses for genetic social effects from selection for direct effects of 0.93, 1.89, -2.19, and 3.52 g for lines 1 to 4.     

托佩克(Topigs)种猪日增重与背标厚选择效果分析

文章对托佩克(Topigs)3个品系5年的日增重和背膘厚性状测定数据作选择效果分析.在对两个性状作遗传参数计算的基础上,明确了这两个性状的遗传力为中等以上,3个品系稍有差异(A系日增重和背膘厚遗传力分别为0.68和0.61;B系分别为0.67和0.48,E系分别为0.48和0.38),同一性状存在阶段性的差别.以此为基...     

Results from nine generations of selection for increased litter size in swine

Direct selection for increased litter size was done for nine generations. The select line consisted of approximately 15 sires and 60 dams per generation, and selection was based on estimated breeding values for number of live pigs. A control line of approximately 10 sires and 30 dams was maintained with stabilizing selection. Heritabilities estimated in the select line using restricted maximal likelihood procedures, daughter-dam regression within sires, and half-sib analysis were 0.01, 0.04, and 0.00 for number of pigs born alive (NBA) and 0.02, 0.16, and 0.00 for total born per litter (TB). Corresponding estimates for the control line were 0.01, 0.06, and 0.23 and 0.02, 0.07, and 0.09 for NBA and TB, respectively. Realized heritabilities for NBA from multiple regression were 0.09 +/- 0.08 in the select line and 0.11 +/- 0.166 in the control line. Heritability estimated from regression of differences in response between lines on differences in cumulative selection differentials was 0.13 +/- 0.07. At Generation 9, litter sizes, estimated breeding values, and cumulative selection differentials were 0.86 (P < 0.05), 0.63 (P < 0.01), and 9.05 (P < 0.01) pigs larger for the select line than for the control line. Phenotypic differences between lines for TB, adjusted backfat (BF), and days to 104 kg (DAYS) were not significant. Genetic trends in the select line were 0.053 +/- 0.002 pigs/yr for NBA, 0.054 +/- 0.013 mm/yr for BF, and 0.398 +/- 0.110 d/yr for DAYS. Corresponding phenotypic trends were 0.145 +/- 0.051 pigs/yr, -0.012 +/- 0.089 mm per yr, and 0.307 +/- 0.278 d/yr, respectively. Genetic trends in the control line were -0.026 +/- 0.004 pigs/yr for NBA, 0.026 +/- 0.022 mm/yr for BF, and -0.532 +/- 0.182 d/yr for DAYS. Corresponding phenotypic trends were 0.001 +/- 0.085 pigs/yr, -0.043 +/- 0.147 mm/yr, and -0.519 +/- 0.462 d/yr, respectively. Litter size can be increased by direct selection using breeding values estimated from an animal model, in conjunction with rearing selected gilts in litters of 10 pigs or less.     

Comparison of genetic improvement for litter size at birth by direct and indirect selection in swine herd

Responses to selection for number of piglets born alive (NBA) by the total number of piglets born (TNB), the NBA, and the NBA plus number of piglets born dead (NBD) were compared using the accuracy of selection and expected genetic gain calculated from the selection index with family information and the real response to selection, using data generated by Monte Carlo computer simulation. The accuracy of selection for NBA selected by TNB was higher than that by NBA only if the genetic correlation between TNB and NBA was close to 1.0, or the value of heritability for the TNB was much larger than that for the NBA. The accuracy of selection for the NBA selected by the combination of the TNB and the NBA was generally highest in the three selection methods in each family structure. Selection by the TNB resulted in the greatest expected genetic gain for the TNB among the selection methods. In the best linear unbiased prediction (BLUP) selection, the genetic gain for the NBA accumulated by the NBA tended to be similar to that accumulated by the combination of the NBA and the NBD, and both genetic gains at generation 10 were significantly larger than that by the TNB (P < 0.001). The accumulated responses selected by the two‐trait animal model BLUP estimated from genetic parameters with errors were similar to those estimated from the true parameters, and there was no significant difference between them. These results indicate that selection by the NBA or by the NBA and the NBD gives more genetic improvement in the NBA than that by the TNB.     

Effect of selection for backfat thickness in swine on fetal skeletal muscle metabolism

At 110 d of gestation, fetuses were removed from sows selected for high (obese) or for low (lean) backfat thickness. The body weights of lean (1,031 +/- 64 g) and obese (864 +/- 55 g) fetuses were not significantly different. Analysis of muscle composition and of in vitro metabolic characteristics was conducted on the biceps femoris muscle. The percentage of dry weight, protein and glycogen was greater in the muscle of obese fetuses than in the muscle of lean fetuses (P less than .01, P less than .05, and P less than .05, respectively). Percentage of muscle triglyceride was similar (P greater than .05) between the two phenotypes. The rate of glucose oxidation to CO2 tended to be greater (P less than .07) and the rate of lactate production was lower (P less than .05) in the muscle from obese fetuses than in the muscle from lean fetuses. The rates of leucine oxidation to CO2 and of palmitate oxidation to CO2 did not differ between phenotypes. The rate of alpha-ketoisocaproate release from the muscle of obese fetuses was greater (P less than .05) than from that of the lean fetuses. The rate of release of alanine and of glutamine plus glutamate did not differ between phenotypes. The rate of esterification of palmitate did not differ between phenotypes. It was concluded that abnormalities in glucose metabolism and in the partitioning of leucine between oxidation and release as the keto acid existed at 110 d of gestation in the muscle of obese fetuses. Any relation between these differences and ultimate differences in carcass composition were not evident.     

Responses in ovulation rate, embryonal survival, and litter traits in swine to 14 generations of selection to increase litter size

Eleven generations of selection for increased index of ovulation rate and embryonal survival rate, followed by three generations of selection for litter size, were practiced. Laparotomy was used to count corpora lutea and fetuses at 50 d of gestation. High-indexing gilts, approximately 30%, were farrowed. Sons of dams in the upper 10% of the distribution were selected. Selection from Generations 12 to 14 was for increased number of fully formed pigs; replacements were from the largest 25% of the litters. A randomly selected control line was maintained. Responses at Generation 11 were approximately 7.4 ova and 3.8 fetuses at 50 d of gestation (P < .01) and 2.3 fully formed pigs (P < .01) and 1.1 live pigs at birth (P < .05). Responses at Generation 14 were three fully formed pigs (P < .01) and 1.4 live pigs (P < .05) per litter. Number of pigs weaned declined (P < .05) in the index line. Total litter weight weaned did not change significantly. Ovulation rate and number of fetuses had positive genetic correlations with number of stillborn pigs per litter. Significantly greater rate of inbreeding and increased litter size at 50 d of gestation in the select line may have contributed to greater fetal losses in late gestation, greater number of stillborn pigs, and lighter pigs at birth, leading to lower preweaning viability. Heritabilities of traits were between 8 and 25%. Genetic improvement programs should emphasize live-born pigs and perhaps weight of live-born pigs because of undesirable genetic relationships of ovulation rate and number of fetuses with numbers of stillborn and mummified pigs and because birth weight decreased as litter size increased.     

Direct responses to selection for increased litter size, decreased age at puberty, or random selection following selection for ovulation rate in swine

Nine generations of selection for high ovulation rate were followed by two generations of random selection and then eight generations of selection for increased litter size at birth, decreased age at puberty, or continued random selection in the high ovulation rate line. A control line was maintained with random selection. Line means were regressed on generation number and on cumulative selection differentials to estimate responses to selection and realized heritabilities. Genetic parameters also were estimated by mixed-model procedures, and genetic trends were estimated with an animal model. Response to selection for ovulation rate was about 3.7 eggs. Response in litter size to selection for ovulation rate was .089 +/- .058 pigs per generation. Average differences between the high ovulation rate and control lines over generations 10 to 20 were 2.86 corpora lutea and .74 pigs (P less than .05). The regression estimate of total response to selection for litter size was 1.06 pigs per litter (P less than .01), and the realized heritability was .15 +/- .05. When the animal model was used, the estimate of response was .48 pigs per litter. Total response in litter size to selection for ovulation rate and then litter size was estimated to be 1.8 and 1.4 pigs by the two methods. Total response to selection for decreased age at puberty was estimated to be -15.7 d (P less than .01) when data were analyzed by regression (realized heritability of .25 +/- .05) and -17.1 d using the animal model. No changes in litter size occurred in the line selected for decreased age at puberty. Analyses by regression methods and mixed-model procedures gave similar estimates of responses and very similar estimates of heritabilities.     

Genetic and environmental trends for litter size in swine

Best linear unbiased predictors (BLUP) of breeding values for additive direct and additive maternal genetic effects were estimated from 3,944 purebred Yorkshire and Landrace first-parity litters recorded on the Quebec Record of Performance Sow Productivity Program and born between 1977 and 1987. Breeding values for gilts, dams, and sires were estimated using an individual animal model for measures of litter size of total number born (NOBN), number born alive (NOBA), and number weaned (NOWN). Environmental trends were estimated from average herd-year solutions, and genetic trends were estimated by regression of estimated breeding value on year of birth. Environmental trends were positive for all traits in both breeds but were significant only for NOWN in Landrace (.051 +/- .021 pigs/yr). Genetic trends were very small but were mainly negative for direct breeding value and combined direct and maternal breeding value. Significant estimates of genetic trends (P less than .05) were observed only within the Yorkshire breed, and these ranged from -.012 +/- .004 to .004 +/- .002 pigs/yr.     

Effects of temperature on the performance of finishing swine: I. Effects of a hot, diurnal temperature on average daily gain, feed intake, and feed efficiency

Ninety-six crossbred barrows and gilts weighing 90 +/- .67 kg were used during a 21-d study to determine the effects of a hot, diurnal temperature (H; 22.5 to 35 degrees C) compared with a constant, thermoneutral temperature (TN; 20 degrees C) and the effects of sex (barrows vs gilts) on performance. A secondary objective included the determination of weight loss as a result of a 24-h fast immediately after the 21-d feeding study of commingled vs not commingled hogs of both environmental treatments (TN and H). Pigs housed in the hot, diurnal temperature gained 16.3% more slowly (P less than .001;.77 vs. .92 kg/d) than those in the constant, thermoneutral environment. Feed intake (FI) for the H pigs was 10.9% less (P less than .001; 3.01 vs 3.38 kg/d) than that for the TN pigs. The H pigs gained 17.6 g/d less and consumed 43.5 g/d less feed for every C degrees above 20 degrees C; however, no differences were observed for feed efficiency (F/G; 3.86 vs 4.19 kg for the TN and H pigs, respectively). Average daily gain and feed/gain (F/G) were not affected by sex. Likewise, no significant interactions of temperature x sex were observed for ADG, FI, or F/G. Weight loss (shrinkage) during the 24-h fast was not affected by commingling; however, the H pigs lost 17.5% more weight (P less than .05) than the TN pigs (3.82 vs 3.25%, respectively).     

Effects of temperature on the performance of finishing swine: II. Effects of a cold, diurnal temperature on average daily gain, feed intake, and feed efficiency

Forty-eight crossbred barrows and gilts weighing 84.5 +/- .33 kg were used during two 21-d trials to investigate the effects of a cold, diurnal temperature (CD; -5.0 to 8.0 degrees C) compared with a constant, thermoneutral temperature (TN; 20 degrees C) and the effects of sex (barrows vs gilts) on performance. A second objective was to determine shrinkage as a result of a 24-h fast immediately after the 21-d study of hogs commingled vs those not commingled for both environmental treatments (CD vs TN). Pigs housed in the CD chamber gained 27.2% more slowly (P less than .001; 75 vs 1.03 kg/d) than those in the TN environment and consumed 5.7% more feed (P less than .05; 3.88 vs 3.67 kg/d). The lower ADG and higher feed intake (FI) exhibited by the CD pigs resulted in poorer (P less than .05) feed efficiency (F/G; 5.33 vs 3.73, respectively). A temperature x sex interaction occurred for ADG but not for FI or F/G. Twenty-four-hour shrink for the CD pigs was 16.4% less than for the TN pigs (3.72 vs 4.45%, respectively); however, commingling did not affect shrinkage.     

Effects of animal and supplement characteristics on average daily gain of grazing beef cattle

Effects of animal gender and age, use of a growth stimulant, and supplementation with grain alone or grain plus other substances on ADG by growing beef cattle grazing bermudagrass paddocks with sod-seeded rye, wheat, and ryegrass were determined. Two grazing experiments (Exp. 1: late winter through mid-spring; Exp. 2: late spring through mid-summer) were conducted. Experiment 1 used 96, 12- to 13-mo-old Simmental-cross calves (heifers, 240 kg; steers, 272 kg), half of which were implanted with zeranol. Within each implant treatment, cattle received no supplement or .5% BW (DM) of ground corn alone or plus a mix of protein meals, zinc sulfate, thiamin-HCl, or salt. Daily gain was higher (P less than .05) with than without supplementation and was similar (P greater than .10) among supplement treatments. In Exp. 2, 96 crossbred beef steers, approximately 7 (230 kg) or 15 mo old (250 kg), were not supplemented (control) or received .5% BW (DM) of ground corn on d 1 to 84 (C-C), corn plus a protein meal mix on d 1 to 84 (CP-CP), corn on d 43 to 84 (O-C), corn plus the protein meal mix on d 43 to 84 (O-CP), or corn on d 1 to 42 and corn plus the protein meal mix on d 43 to 84 (C-CP). Daily gain on d 1 to 84 was affected (P less than .05) by supplement, age, implant, and the supplement x implant interaction (nonimplanted: .37, .56, .68, .40, .49, and .49; implanted: .37, .62, .54, .49, .70, and .71 kg for control, C-C, CP-CP, O-C, O-CP, and C-CP, respectively).     

Alternative methods of selection for litter size in mice: II. Response to thirteen generations of selection

Selection was conducted on an index of components of litter size (I = 1.21 x ovulation rate + 9.05 x ova success; ovulation rate measured by number of corpora lutea and ova success measured as number of pups born + number of corpora lutea), on uterine capacity (measured as number of pups born to unilaterally ovariectomized dams) and on litter size concurrent with an unselected control for 13 generations. Selection criteria (IX = index, UT = uterine capacity, LS = litter size and LC = control) were applied in each of three replicates. In an evaluation after five generations, IX and LS each exceeded LC by about .5 pups, with no response in UT. After 13 generations, mean ovulation rate, ova success and litter size (measured as number of fetuses at 17 d gestation in intact females) were, for IX, 14.25, .84, 11.95; for LS, 14.15, .82, 11.64; for UT, 12.61, .86, 10.77; and for LC, 12.27, .82, 9.98. The regression of number born (litter size in IX, LS and LC; uterine capacity with only a functional left uterine horn in UT) on cumulative selection differential across 13 generations was .12 +/- .01, .09 +/- .02 and .08 +/- .02 for IX, LS and UT, respectively. The regression of breeding value for litter size on each selection criterion, estimated as response in the generation-13 evaluation divided by cumulative selection differential, was .11 +/- .02, .08 +/- .01 and .05 +/- .03 for IX, LS and UT, respectively. Regression of response in number born on generation number was .17 +/- .01, .15 +/- .04 and .10 +/- .02 for IX, LS and UT, respectively. Selection in IX was promising relative to LS, and selection in UT changed number born.     

Effects of correcting missing daily feed intake values on the genetic parameters and estimated breeding values for feeding traits in pigs

Daily feed intake (DFI) is an important consideration for improving feed efficiency, but measurements using electronic feeder systems contain many missing and incorrect values. Therefore, we evaluated three methods for correcting missing DFI data (quadratic, orthogonal polynomial, and locally weighted (Loess) regression equations) and assessed the effects of these missing values on the genetic parameters and the estimated breeding values (EBV) for feeding traits. DFI records were obtained from 1622 Duroc pigs, comprising 902 individuals without missing DFI and 720 individuals with missing DFI. The Loess equation was the most suitable method for correcting the missing DFI values in 5–50% randomly deleted datasets among the three equations. Both variance components and heritability for the average DFI (ADFI) did not change because of the missing DFI proportion and Loess correction. In terms of rank correlation and information criteria, Loess correction improved the accuracy of EBV for ADFI compared to randomly deleted cases. These findings indicate that the Loess equation is useful for correcting missing DFI values for individual pigs and that the correction of missing DFI values could be effective for the estimation of breeding values and genetic improvement using EBV for feeding traits.     

Comparison of three models to estimate breeding values for percentage of loin intramuscular fat in Duroc swine

Three selection models were evaluated to compare selection candidate rankings based on EBV and to evaluate subsequent effects of model-derived EBV on the selection differential and expected genetic response in the population. Data were collected from carcass- and ultrasound-derived estimates of loin i.m. fat percent (IMF) in a population of Duroc swine under selection to increase IMF. The models compared were Model 1, a two-trait animal model used in the selection experiment that included ultrasound IMF from all pigs scanned and carcass IMF from pigs slaughtered to estimate breeding values for both carcass (C1) and ultrasound IMF (U1); Model 2, a single-trait animal model that included ultrasound IMF values on all pigs scanned to estimate breeding values for ultrasound IMF (U2); and Model 3, a multiple-trait animal model including carcass IMF from slaughtered pigs and the first three principal components from a total of 10 image parameters averaged across four longitudinal ultrasound images to estimate breeding values for carcass IMF (C3). Rank correlations between breeding value estimates for U1 and C1, U1 and U2, and C1 and C3 were 0.95, 0.97, and 0.92, respectively. Other rank correlations were 0.86 or less. In the selection experiment, approximately the top 10% of boars and 50% of gilts were selected. Selection differentials for pigs in Generation 3 were greatest when ranking pigs based on C1, followed by U1, U2, and C3. In addition, selection differential and estimated response were evaluated when simulating selection of the top 1, 5, and 10% of sires and 50% of dams. Results of this analysis indicated the greatest selection differential was for selection based on C1. The greatest loss in selection differential was found for selection based on C3 when selecting the top 10 and 1% of boars and 50% of gilts. The loss in estimated response when selecting varying percentages of boars and the top 50% of gilts was greatest when selection was based on C3 (16.0 to 25.8%) and least for selection based on U1 (1.3 to 10.9%). Estimated genetic change from selection based on carcass IMF was greater than selection based on ultrasound IMF. Results show that selection based on a combination of ultrasonically predicted IMF and sib carcass IMF produced the greatest selection differentials and should lead to the greatest genetic change.     

Effect of exogenous progesterone and oestrone on litter size in swine

Three hundred sows and gilts on a large commercial unit were divided into four groups according to parity and either injected with corn oil on days 16 and 17 after service, injected with 25 mg progesterone and 12.5 micrograms oestrone on days 16 and 17 after service, or injected with these two hormones on either day 16 or day 17 after service. Animals injected with progesterone and oestrone on both the 16th and 17th days after mating had significantly (P less than 0.05) bigger litters at farrowing. Animals injected on the 17th day or on both the 16th and 17th days after mating had significantly shorter gestation periods (P less than 0.05). The treatments had no effect on the weaning to service interval or on the size of the subsequent litter.     

Effects of selecting for litter size and body weight on the components of litter size in mice

Divergent selection for the total number of young born in the first three litters (TNY-3) and for 6-week body weight in mice resulted in different changes in the components of first litter size.Selection for large TNY-3 increased the number of ova shed while pre-natal losses were not modified. Selection for small TNY-3 caused both a decrease in ovulation rate and an increase in post-implantation mortality. Divergent selection for body weight modified ovulation rate in the same direction as body weight. Embryo survival declined in both lines selected for body weight. That decline was caused by an increase in pre-implantation mortality in the case of the high line, and by an increase in post-implantation mortality in the low line.     

Estimation of variances for gametic effects on litter size in Yorkshire and Landrace swine

The objective of this study was to test for effects of gametic imprinting on litter size in swine by estimating variances for parent-specific gametic effects. Data were 64,047 and 137,009 multiparous records of number born alive for the U.S. Landrace and Yorkshire breeds, respectively. The statistical model included fixed effects of parity number and herd, and random effects of herd-year-season, mate, permanent environment, animal (additive genetic), and either maternal or paternal gametes. A Bayesian approach that used Gibbs sampling to obtain posterior distributions was employed. To aid in the interpretation of results, the Landrace data structure was used to simulate data with and without effects of imprinting. Analyses of the simulated records indicated that the model applied was capable of detecting effects of imprinting when such effects were present. Small, but non-zero, estimates of gametic variances were obtained when no imprinting was simulated. Estimates of the proportion of total variance accounted for by paternally transmitted gametes were 0.8 and 0.9% for Landrace and Yorkshires, respectively. These estimates were different from zero, but were similar to the results observed for data simulated without an imprinting effect. Corresponding results for maternally transmitted gametes were 1.6% for Landrace and 0.8% for Yorkshires. The estimate for Landrace was significantly greater than that observed for Yorkshires and for the simulations without a true effect and suggested the presence of a non-Mendelian genetic influence on litter size. Paternally imprinted genes are a plausible reason for the observed results. Assuming that the effect observed was due to paternal imprinting at a single biallelic locus, the substitution effect of the superior allele could be greater than 0.7 piglets per litter. Identification of a genetic marker for such an allele would be useful in marker-assisted selection of females. Other possible explanations exist for the increased gametic variance in the Landrace breed, but these explanations (such as maternal or cytoplasmic effects) may be less likely than paternal imprinting.