Detecting parent of origin and dominant QTL in a two-generation commercial poultry pedigree using variance component methodology
Genetics Selection Evolutionvolume 41, Article number: 6 (2009)
Variance component QTL methodology was used to analyse three candidate regions on chicken chromosomes 1, 4 and 5 for dominant and parent-of-origin QTL effects. Data were available for bodyweight and conformation score measured at 40 days from a two-generation commercial broiler dam line. One hundred dams were nested in 46 sires with phenotypes and genotypes on 2708 offspring. Linear models were constructed to simultaneously estimate fixed, polygenic and QTL effects. Different genetic models were compared using likelihood ratio test statistics derived from the comparison of full with reduced or null models. Empirical thresholds were derived by permutation analysis.
Dominant QTL were found for bodyweight on chicken chromosome 4 and for bodyweight and conformation score on chicken chromosome 5. Suggestive evidence for a maternally expressed QTL for bodyweight and conformation score was found on chromosome 1 in a region corresponding to orthologous imprinted regions in the human and mouse.
Initial results suggest that variance component analysis can be applied within commercial populations for the direct detection of segregating dominant and parent of origin effects.
Despite intense selection there is evidence to suggest that there is still much variation that might be exploited within commercial populations [1, 2]. The effectiveness of selection procedures utilising genomic information can be increased by correctly identifying the mode of inheritance of desired variants. For example, Hayes and Miller  show that including dominance effects in mate selection can be a powerful tool for exploiting previously untapped genetic variation while Dekkers and Chakraborty  discuss maximization of crossbred performance by incorporating information from overdominant QTL.
Historically, much of the success in commercial poultry breeding and many other agricultural species has relied on utilizing heterosis and reciprocal effects [5–8], yet the underlying genetic architecture is still not clear. It appears that both maternal effects and dominant or over-dominant genes play a role . Tuiskula-Haavisto and Vilkki  suggest that there is also recent evidence for the role of parentally imprinted mechanisms in poultry to explain the underlying mechanism for reciprocal effects.
Despite increasing evidence for parent of origin effects in crosses between divergent lines of poultry, imprinting in poultry remains a contentious issue.
Genomic imprinting affects many mammalian genes  and is brought about by epigenetic instructions or imprints that are laid down in the parental germ cells . Imprinting is most prevalent in foetal development and until recently was considered best described by the parental conflict hypothesis . In viviparous animals this occurs where the male exerts selection pressure for offspring to maximise use of maternal resources whereas the female limits this allocation of resources to preserve herself and future offspring. As there is no apparent parental conflict, the presence of imprinting was not thought to occur in oviparous species. Furthermore, IGF2 has been shown to be imprinted and expressed from paternal allele in man rabbit, mice, pig, and sheep [14, 15], but not in the chicken . There is, however, recent evidence for imprinted genes in birds and lower vertebrates and for shared orthologues with mammalian imprinted genes [17, 18]. Different species may also have species specific imprinted genes . Current theory suggests that the evolution of imprinted genes is a dynamic step-wise process with orthologues present on separate chromosomes before imprinting arose. These conserved orthologues were selected during vertebrate evolution becoming imprinted only as the need arose [18, 20]. Lawton et al.,  show that transcriptional silencing at imprinted loci has evolved along independent trajectories in mammals and marsupials. Imprinted genes are characteristically found in a clustered organization with 80% physically linked with other imprinted genes. These clusters are conserved in mammals, marsupials and flowering plants. . Studies reporting QTL with parent of origin effects in chicken show a similar pattern tending to cluster on a few macrochromosomes with 78% of imprinted gene orthologues residing on chicken chromosomes 1, 3, and 5 [10, 18].
Both dominant and imprinted QTL effects have been identified in poultry for economically important production and disease resistance traits. Ikeobi et al.,  found that 1/3 of QTL found for fat related traits in a broiler-layer cross showed dominance effects; Yonash et al.,  found both partial and overdominance QTL effects for resistance to Marek's disease, while Kerje et al.,  and Tuiskula-Haavisto et al.,  report dominant effects for egg production traits. Parent of origin effects in poultry are reviewed by Tuiskula-Haavisto et al.,  and have been found for bodyweight, carcass and egg production traits [26–28].
All of these studies have involved crosses between lines or divergent populations, reviewed by Hocking  and Abasht et al.,. Detection of QTL effects, however, within model organisms or experimental populations is costly and potentially of limited relevance to populations under selection. It is of much greater benefit to directly explore QTL segregating within commercial populations. A variance component or pedigree based approach can be applied to map QTL directly within the population under selection and by simple extension of genetic models can potentially also be used to dissect the mode of inheritance at the QTL. Here we use a variance component approach to look for dominant and imprinted QTL associated with bodyweight and conformation score measured at 40 days in a two-generation commercial broiler population.
Phenotypes on conformation score and bodyweight, both measured at 40 days, were available for a commercial broiler dam line from Cobb Breeding Company Ltd. Conformation score is a subjective measure of fleshiness scored from 1–5 and was treated as normally distributed. A two-generation pedigree was available with a total of 2708 offspring with phenotypes and genotypes for markers in candidate QTL regions on chicken chromosomes 1, 4 and 5. Candidate regions were based on a previous three generation study of the Cobb population . Forty-six sires were mated to 100 dams with an average of two dams per sire, 59 half sibs per sire and 27 full sibs per dam. The number of progeny per sires and dam ranged from 9 to 149 and 14 to 44, respectively. Birds were genotyped for markers spaced approximately every 16, 14 and 8 cM on chromosomes 1, 4, and 5, respectively. Markers were selected from the consensus linkage map . Linkage maps were estimated using CriMap  and linkage groups corresponded to the consensus map at approximately 128–205 cM, 75 – 182 cM, and 57–104 cM for chicken chromosomes 1, 4 and 5 respectively. Marker distances and consensus map positions are given in additional file 1. Progeny were from two flocks across 17 hatch weeks. Fixed effects of sex, age of dam, and hatch within flock were fitted. Summary statistics and heritabilities can be found in Table 1. The correlation between the two traits was 0.34 (0.03). Further details can be found in Rowe et al. .
Statistical genetic models for variance component analysis
Following a two-step approach similar to that described by George et al. , identical by descent (IBD) coefficients were estimated for all relationships in the pedigree to calculate the covariance matrices for the QTL effects, which were subsequently used in a linear mixed model.
The G, G M , G P and D are the appropriate relationship matrices used to model the additive, maternal, paternal and dominant QTL effects at each position tested. They are conditional on flanking marker information and therefore unique for each position evaluated for a QTL. Here the matrices were calculated every 5 cM.
It can be shown that these relationship matrices are easily estimated from the gametic IBD matrix, a 2n × 2n matrix containing the probability of identity of descent between any of the two gametes of an individual with the gametes of the remaining individuals in the pedigree . Where P1 is the paternally derived allele at a locus and P2 is the maternally derived allele elements of the gametic IBD matrix for individuals i and j at a single locus are . From this the additive covariance between i and j is r ij = 1/2(P11 + P12 + P21 + P22) and the covariance due to dominance i.e. the inheritance of two alleles identical by descent is u ij = P11P22 + P12P21 . The probability of individuals i and j sharing paternal or maternal QTL alleles IBD is simply P11 or P22 respectively.
In contrast to George et al.  who used a Monte-Carlo method, the gametic IBD matrix was estimated with the recursive method of Pong-Wong et al.,  (software available on request from the author), which uses the closest fully informative or phase-known flanking marker to estimate the IBD at the putative QTL. Variance components for each model were estimated using REML  implemented in the ASReml package .
The statistical models used were:
(1) y = Xβ + Zu + Wc + e (null or polygenic)
(2) y = Xβ + Zu + Wc + Za + e (additive QTL)
(3) y = Xβ + Zu + Wc + Za + Zd +e (additive QTL + dominant QTL)
(4) y = Xβ + Zu + Wc + Z m m + Z p p +e (maternal QTL + paternal QTL)
(5) y = Xβ + Zu + Wc + Z p p +e (paternal QTL)
(6) y = Xβ + Zu + Wc + Z m m +e (maternal QTL)
where y is a vector of phenotypic observations, β is a vector of fixed effects, u, a, d, m, p, c and e are vectors of random additive polygenic effects, additive and dominance QTL effects, maternal and paternal QTL effects, maternal effects and residuals, respectively. X, Z, W, Z m , and Z p are incidence matrices relating to fixed and random genetic, direct maternal, maternally expressed, and paternally expressed QTL effects, respectively.
Variances for polygenic and QTL effects are distributed as follows: var(u) =A σ2a, Var(a) = G σ2q, Var(d) = D σ2d, Var(m) = G M σ2m, Var(p) = G P σ2p, var(e) = I σ2e. For the non-genetic maternal effect Var(c) = I σ2c. A is the standard additive relationship matrix based on pedigree data only and the relationship matrices G, G M , G P and D for a given QTL position are calculated from the gametic IBD matrix as outlined by Liu et al.,.
A test statistic for a given location was obtained by comparing the likelihood of the full versus the reduced model. Twice the difference between the log likelihood of the full versus the reduced model was used as a log likelihood ratio test (LRT). Permutation was used to set significant thresholds. For linkage group-wise test statistics, genotypes were permuted within dam families to remove associations with IBD status and phenotype. Because permutation was done within dam families, i.e. sibs swap genotypes but retain phenotypes, the A matrix and therefore the estimated polygenic variance remained the same. After each permutation, analyses for all models were repeated for every test position along the chromosome and the highest test statistic was recorded. After 1000 permutations the test statistics were ranked and the 95th percentile used for a linkage group-wise 5% type 1 error rate. Separate permutation analyses were carried out for each trait. Permutation analysis for all three chromosomes was similar so thresholds were set using the results from chromosome 4 as this is the linkage group with the most tests. In each case the highest test statistic for each model was recorded regardless of position. Empirical thresholds for each test are given in Table 2. For plotting purposes test statistics for each position were converted to rank by comparison to the results of the permutation analysis, and the rank subsequently divided by 10. For example, a test statistic corresponding to the 950th ranked value from the permutation analysis was plotted with a value of 95, corresponding to a 5% type 1 error.
Detection of dominant QTL effects
To detect dominant QTL effects, three tests were carried out:
(i) add, comparing the additive QTL model (2) versus the null model (1) to test significance of the QTL variance component under a purely additive model;
(ii) addom, comparing the additive QTL + dominance QTL model (3) versus the null model (1) to test significance of QTL variance components under a model including additive and dominance effects;
(iii) dom, comparing the additive QTL + dominance QTL model (3) vs. the additive QTL model (2) to test the significance of the dominance variance component.
Tests (i) and (ii) are used in the initial search for the QTL whereas test (iii) is applied subsequently to test specifically for the dominance component. The dom test was applied at all positions regardless of significance of other tests.
Parent of origin effects
Initially QTL can be searched for using additive (add), pat + mat or single parental models (mat or pat). To test for imprinting four tests were carried out at each position:
(i) pat + mat, comparing the paternal QTL + maternal QTL model (4) vs. the null model (1) to test the significance of an additive QTL whilst allowing the maternal and paternal components to vary;
(ii) imp, comparing the pat + mat model (4) vs. the add model (2) to test whether the additive effect was better explained by allowing different parental contributions;
(iii) patvfull, comparing the paternal QTL model pat (5) vs. the pat + mat model to test for contribution of a paternally inherited QTL to the QTL variance;
(iv) matvfull, comparing the paternal QTL model mat (6) vs. the pat + mat model to test for contribution of a maternally inherited QTL to the QTL variance.
Again, all tests were carried out at all positions regardless of significance of other tests. Following Hanson et al.,  under an additive model both parents contribute equally whereas for an imprinted QTL only one parent is expected to show expression. For example, for a maternally expressed QTL the expectation is that the patvfull test is significant and the matvfull test is not significant. For non-imprinted QTL the expectation is that both tests are significant because there is expression from both parents.
Common environment effects are often, at least partially, confounded with dominance and imprinting as shown by Rowe et al.,  thus common family environment or 'dam' effects were included in all models.
Table 1 gives heritabilities for the two traits. These are low, probably due to selection of the parents . Because heritability estimates are based on the contrast of between and within family variance and QTL variance is based mainly on within family variance low trait heritability is not expected to affect QTL detection.
Additive and dominant QTL effects
Figure 1 shows QTL effects under additive and dominant QTL models for bodyweight and conformation score. There were chromosome-wide significant dominant QTL effects for conformation score on chromosomes 4 and 5. These effects were considerable, explaining 6.2 and 4.5% of the phenotypic variance, respectively. Table 3 shows that the dominant QTL explains all of the QTL variance (i.e. the estimated additive effect of the QTL is zero when a model with both an additive and dominant QTL effect is fitted).
Parent-of-origin QTL effects
Figure 2 shows rank of test statistics when compared to permutation analysis for bodyweight on chromosomes 1, 4 and 5. Figure 1 shows that there was not significant evidence for a purely additive QTL at the beginning of the chromosome 1. Figure 2, however, shows that the pat + mat model is significantly better than the add model and there is evidence for a maternally expressed QTL on chromosome 1. Table 4 also shows that the patvfull test is significant whereas the matvfull test is not indicating maternal expression. Furthermore, all of the QTL variance is explained by the maternal QTL (Table 5).
Figure 3 shows test statistics for conformation score on chromosomes 1, 4 and 5. For chromosomes 1 and 5 there is some evidence for a maternally expressed QTL affecting conformation score although the imp test is only significant for chromosome 1. Chromosome 4 has two linkage peaks, however neither reaches significance.
Dominant and parentally expressed QTL effects were distinct from one another and do not appear to be confounded therefore they are discussed separately.
There is suggestive evidence for a maternally expressed QTL on chromosome 1 for both weight and conformation score associated with marker interval ADL0307-LEI0068, a region orthologous with imprinted regions in the mouse and human associated with Prader-Willi/Angelman syndrome . This region of chromosome 1, corresponding to approximately 128 to 151 cM on the consensus map, is within a marker interval associated with many fat and carcass traits in chickens [22, 24, 30, 43, 44]. Furthermore, McElroy et al.,  and Tuiskula-Haavisto et al.,  both find maternally expressed QTL within the same marker bracket associated with egg production. Sharman et al  find imprinted effects for skeletal traits at 135 cM on chromosome 1. Tuiskula-Haavisto et al.,  also find a paternally expressed QTL associated with age at first egg in the same marker interval as the putative paternally expressed effect seen here for conformation score.
When a common environment or dam effect was omitted from the model (results not shown) evidence, in particular, for the maternally expressed QTL on chromosome 1 increased. The mat + pat and imp tests also reached significance if a dam effect was not accounted for. This is possibly due to confounding of effects i.e. common environment can give spurious variance at the QTL and further highlights the importance of fitting maternal effects to avoid spurious detection of QTL. De Koning et al  found significant additive effects for bodyweight and conformation and a strong dam effect associated with this region using a three generation design from the same population. This could indicate that a strong component of the effect on chromosome 1 associated with bodyweight and conformation score comes from maternally influenced egg traits. Maternal influence on fresh egg weight and subsequent bodyweight, particularly early growth is well documented [45–47]. Kerje et al  report a strong correlation between egg weight and adult bodyweight (r = 0.62, 0.0001) and a QTL for growth at the beginning of chromosome 1 explaining half the phenotypic variation seen in egg weight.
There appear to be two separate QTL segregating for bodyweight and conformation score on chromosome 4. For bodyweight there is an additive QTL in the region of ADL0266 - LEI0076 as found by Kerje et al.,  and Jacobsson et al., . There is greater evidence for this from the paternal analysis. Although the paternal QTL appears to explain most of the additive variance there is insufficient evidence for imprinting i.e. the test of the pat + mat model versus an additive model is not significant. For conformation, a dominant and potentially over-dominant QTL explaining all of the QTL variance maps to around 80–118 cM on the consensus map. Yonash et al.,  find partial and overdominance for QTL affecting resistance to Marek's disease in this marker bracket. Although Ikeobi et al  find many dominant effects for carcass trait QTL, they find the QTL on chromosome 4 tends to behave additively as a single locus affecting many traits. Sharman et al  report QTL for many traits associated with skeletal traits on chromosome 4 including a dominant QTL associated with tibial marrow diameter at ADL0266-ROS0024.
On chromosome 5 there appear to be dominant effects for bodyweight and conformation traits. Although the test for dominance (dom) is significant for bodyweight the actual QTL does not reach linkage group-wise significance. Ikeobi et al.,  also found modest dominance effects for growth traits in this region. For conformation score, there is evidence for most of chromosome 5 for a significant dominant QTL and maternal expression at the end of the linkage group. Abasht et al.,  also find a maternal sex interaction with fat traits in this marker bracket. Chromosome 5 has been associated with many paternally expressed traits [27, 28] and although the linkage group does not span the region, the first marker interval is close to a conserved gene cluster of twelve imprinted gene orthologues shown to replicate asynchronously. Despite this, here we see no evidence for paternal imprinting on chromosome 5. Ikeobi et al., find many QTL for traits associated with weight and carcass composition in this region although little dominance and no imprinting.
Given that we are only using a two-generation pedigree we have insufficient evidence to confirm that these are truly imprinted effects, only that statistically there is evidence for uniparental expression. Heuven et al.,  show that spurious imprinted effects can be detected due to differences between the number of QTL alleles or haplotypes segregating in sires and dams. This can occur for a number of reasons, for example; too few sires or dams included in the analysis, different genetic backgrounds leading to different QTL allele frequencies in sires and dams and or differing amounts of LD generated between QTL alleles and markers.
To ensure information on a putative QTL is available from both parents there is a requirement for enough sires and dams to ensure segregation together with enough offspring to detect QTL. Furthermore the QTL allele frequency should be roughly equal in sires and dams. Here, these requirements are satisfied by using a large number of sires and dams. Furthermore, because the analysis took place within a broiler dam line, i.e. sires and dams have the same genetic background, neither differing allele frequencies due to parental origins or sampling issues are likely causes of spurious imprinting. Marker allele frequencies are not significantly different between sires and dams and an average of 24 sires and 25 dams are informative at any given marker (results not shown).
It is possible that differences in LD between the marker and QTL alleles might occur if parents originate from different populations, however again this is not the case here and furthermore, marker spacing makes it unlikely that strong LD in one sex could have caused differences in variation as it has been shown that linkage disequilibrium in commercial poultry populations rarely exceeds 1 or 2 cM . Using simulation, Tuiskula-Haavisto et al.,  also concluded that segregation differences are an unlikely source of spurious parent-of-origin effects.
A further source of error might be spurious detection of maternally expressed QTL due to common maternal environment, here common environment is fitted within the linear model. Finally, it is feasible that there are many QTL causing a complex inheritance pattern although again due to sires and dams coming from the same lines it is unlikely that different QTL would be segregating. It would be difficult to test this using the current structure due to the complexity of the analysis as it is unlikely that the extra number of variance components added could be successfully estimated. It is also unlikely that enough information could be derived from the marker spacing to estimate multiple QTL within discrete confidence intervals.
Further evidence for the results found here is that imprinted effects on chromosome 1 were found in regions previously identified as parentally expressed in poultry and orthologous with genome-imprinted regions in humans and mice.
Testing many models at each position raises its own multiple testing issues, one strategy might be to only carry out subsequent testing after identifying a significant additive QTL. This, however, can lead to QTL being missed due to the use of an inappropriate model. When testing for dominance the dominant QTL on chromosome 4 would not have been detected under an additive model.
Similarly for parentally expressed QTL, it follows that the contrast may not be greatest at the highest test statistic for the pat + mat or add models but at the highest test statistic for the individual parental QTL i.e. mat or pat. For example, on chromosome 5 the greatest evidence for a maternal QTL and for the imp test is not at the same position as the highest test statistic for a search under the pat + mat model versus null. On chromosome 1, there is a maternal QTL and the imp test is significant, however the pat + mat model is not. The pat + mat model versus null is perhaps diluted by the non expression from the imprinted parent as it is explaining the same amount of variation with an extra degree of freedom. Here we also find that a bodyweight QTL on chromosome 4 could be declared as paternally expressed based upon separate parental QTL models but there is insufficient evidence when comparing a Mendelian versus a pat + mat or imprinted model. It is difficult to know whether this is due to information source, or perhaps too stringent a threshold on the imp test or too lenient on the pat test.
A large dominant and potentially over-dominant QTL for conformation score is segregating on chicken chromosome 4. This QTL is also detected under an additive model. However, the additive variance becomes zero in a model that also fits a dominance component. There is also evidence for dominant QTL affecting bodyweight and conformation on chromosome 5. There is suggestive evidence for a paternally imprinted or maternally expressed QTL affecting bodyweight and conformation score on chromosome 1 in a region orthologous with human and mouse imprinted regions and close to previously reported imprinted QTL affecting bodyweight and maternal traits in poultry. Initial results suggest that variance component analysis can be applied within commercial populations for the direct detection of segregating dominant and parent of origin effects.
De Koning DJ, Haley CS, Windsor D, Hocking PM, Griffin H, Morris A, Vincent J, Burt DW: Segregation of QTL for production traits in commercial meat- type chickens. Genet Res. 2004, 83: 211-220. 10.1017/S0016672304006846.
Andersson L, Georges M: Domestic-animal genomics: deciphering the genetics of complex traits. Nat Rev Genet. 2004, 5: 202-212. 10.1038/nrg1294.
Hayes BJ, Miller SP: Mate selection strategies to exploit across- and within-breed dominance variation. J Anim Breed Genet. 2000, 117: 347-359. 10.1046/j.1439-0388.2000.00252.x.
Dekkers JCM, Chakraborty R: Optimizing purebred selection for crossbred performance using QTL with different degrees of dominance. Genet Sel Evol. 2004, 36: 297-324. 10.1051/gse:2004003.
Fairfull RW, Gowe RS, Emsley JA: Diallel cross of six long-term selected leghorn strains with emphasis on heterosis and reciprocal effects. Br Poult Sci. 1983, 24: 133-158. 10.1080/00071668308416724.
Liu G, Dunnington EA, Siegel PB: Growth related traits in body weight selected lines and their crosses reared under different nutritional regimens. Br Poult Sci. 1995, 36: 209-219. 10.1080/00071669508417770.
Nestor KE, Anderson JW, Velleman SG: Genetic variation in pure lines and crosses of large-bodied turkey lines. 3. Growth-related measurements on live birds. Poult Sci. 2005, 84: 1341-1346.
Marks HL: Heterosis and overdominance following long-term selection for body weight in Japanese quail. Poult Sci. 1995, 74: 1730-1744.
Fairfull RW: Heterosis. 1990, New York: Elsevier Science, 913-933.
Tuiskula-Haavisto M, Vilkki J: Parent-of-origin specific QTL–a possibility towards understanding reciprocal effects in chicken and the origin of imprinting. Cytogenet Genome Res. 2007, 117: 305-312. 10.1159/000103192.
Morison IM, Ramsay JP, Spencer HG: A census of mammalian imprinting. Trends Genet. 2005, 21: 457-465. 10.1016/j.tig.2005.06.008.
Reik W, Walter J: Evolution of imprinting mechanisms: the battle of the sexes begins in the zygote. Nat Genet. 2001, 27: 255-256. 10.1038/85804.
Moore T, Haig D: Genomic imprinting in mammalian development: a parental tug-of-war. Trends Genet. 1991, 7: 45-49.
Nezer C, Moreau L, Wagenaar D, Georges M: Results of a whole genome scan targeting QTL for growth and carcass traits in a Pietrain × Large White intercross. Genet Sel Evol. 2002, 34: 371-387. 10.1051/gse:2002013.
Morison IM, Paton CJ, Cleverley SD: The imprinted gene and parent-of-origin effect database. Nucleic Acids Res. 2001, 29: 275-276. 10.1093/nar/29.1.275.
Yokomine T, Kuroiwa A, Tanaka K, Tsudzuki M, Matsuda Y, Sasaki H: Sequence polymorphisms, allelic expression status and chromosome locations of the chicken IGF2 and MPR1 genes. Cytogenet Cell Genet. 2001, 93: 109-113. 10.1159/000056960.
Dunzinger U, Haaf T, Zechner U: Conserved synteny of mammalian imprinted genes in chicken, frog, and fish genomes. Cytogenet Genome Res. 2007, 117: 78-85. 10.1159/000103167.
Dunzinger U, Nanda I, Schmid M, Haaf T, Zechner U: Chicken orthologues of mammalian imprinted genes are clustered on macrochromosomes and replicate asynchronously. Trends Genet. 2005, 21: 488-492. 10.1016/j.tig.2005.07.004.
Okamura K, Ito T: Lessons from comparative analysis of species-specific imprinted genes. Cytogenet Genome Res. 2006, 113: 159-164. 10.1159/000090828.
Edwards CA, Rens W, Clarke O, Mungall AJ, Hore T, Graves JA, Dunham I, Ferguson-Smith AC, Ferguson-Smith MA: The evolution of imprinting: chromosomal mapping of orthologues of mammalian imprinted domains in monotreme and marsupial mammals. BMC Evol Biol. 2007, 7: 157-10.1186/1471-2148-7-157.
Lawton BR, Carone BR, Obergfell CJ, Ferreri GC, Gondolphi CM, Vandeberg JL, Imumorin I, O'Neill RJ, O'Neill MJ: Genomic imprinting of IGF2 in marsupials is methylation dependent. BMC Genomics. 2008, 9: 205-10.1186/1471-2164-9-205.
Ikeobi CO, Woolliams JA, Morrice DR, Law A, Windsor D, Burt DW, Hocking PM: Quantitative trait loci affecting fatness in the chicken. Anim Genet. 2002, 33: 428-435. 10.1046/j.1365-2052.2002.00911.x.
Yonash N, Bacon LD, Witter RL, Cheng HH: High resolution mapping and identification of new quantitative trait loci (QTL) affecting susceptibility to Marek's disease. Anim Genet. 1999, 30: 126-135. 10.1046/j.1365-2052.1999.00457.x.
Kerje S, Carlborg O, Jacobsson L, Schütz K, Hartmann C, Jensen P, Andersson L: The twofold difference in adult size between the red junglefowl and White Leghorn chickens is largely explained by a limited number of QTLs. Anim Genet. 2003, 34: 264-274. 10.1046/j.1365-2052.2003.01000.x.
Tuiskula-Haavisto M, Honkatukia M, Vilkki J, De Koning DJ, Schulman NF, Maki-Tanila A: Mapping of quantitative trait loci affecting quality and production traits in egg layers. Poult Sci. 2002, 81: 919-927.
Tuiskula-Haavisto M, De Koning DJ, Honkatukia M, Schulman NF, Maki-Tanila A, Vilkki J: Quantitative trait loci with parent-of-origin effects in chicken. Genet Res. 2004, 84: 57-66. 10.1017/S0016672304006950.
Sharman PW, Morrice DR, Law AS, Burt DW, Hocking PM: Quantitative trait loci for bone traits segregating independently of those for growth in an F2 broiler × layer cross. Cytogenet Genome Res. 2007, 117: 296-304. 10.1159/000103191.
McElroy JP, Kim JJ, Harry DE, Brown SR, Dekkers JC, Lamont SJ: Identification of trait loci affecting white meat percentage and other growth and carcass traits in commercial broiler chickens. Poult Sci. 2006, 85: 593-605.
Hocking PM: Review of QTL mapping results in chickens. Worlds Poult Sci J. 2005, 61: 215-226. 10.1079/WPS200461.
Abasht B, Dekkers JC, Lamont SJ: Review of quantitative trait loci identified in the chicken. Poult Sci. 2006, 85: 2079-2096.
Schmid M, Nanda I, Guttenbach M, Steinlein C, Hoehn M, Schartl M, Haaf T, Weigend S, Fries R, Buerstedde JM, Wimmers K, Burt DW, Smith J, A'Hara S, Law A, Griffin DK, Bumstead N, Kaufman J, Thomson PA, Burke T, Groenen MA, Crooijmans RP, Vignal A, Fillon V, Morisson M, Pitel F, Tixier-Boichard M, Ladjali-Mohammedi K, Hillel J, Mäki-Tanila A, Cheng HH, Delany ME, Burnside J, Mizuno S: First report on chicken genes and chromosomes 2000. Cytogenet Cell Genet. 2000, 90: 169-218. 10.1159/000056772.
Green P, Falls K, Crooks S: Documentation for CRI-MAP. Version 2 4. 1990
Rowe SJ, Windsor D, Haley CS, Burt DW, Hocking PM, Griffin H, Vincent J, De Koning DJ: QTL analysis of body weight and conformation score in commercial broiler chickens using variance component and half-sib analyses. Animal Genetics. 2006, 37: 269-272. 10.1111/j.1365-2052.2006.01424.x.
George AW, Visscher PM, Haley CS: Mapping quantitative trait loci in complex pedigrees: A two- step variance component approach. Genetics. 2000, 156: 2081-2092.
Pong-Wong R, George AW, Woolliams JA, Haley CS: A simple and rapid method for calculating identity-by-descent matrices using multiple markers. Genet Sel Evol. 2001, 33: 453-471. 10.1051/gse:2001127.
Patterson HD, Thompson R: Recovery of inter-block information when block sizes are unequal. Biometrika. 1971, 58: 545-554. 10.1093/biomet/58.3.545.
Gilmour AR, Thompson R, Cullis BR: Average Information REML, an efficient algorithm for variance parameter estimation in linear mixed models. Biometrics. 1995, 51: 1440-1450. 10.2307/2533274.
Liu Y, Jansen GB, Lin CY: The covariance between relatives conditional on genetic markers. Genet Sel Evol. 2002, 34: 657-678. 10.1051/gse:2002030.
Hanson RL, Kobes S, Lindsay RS, Knowler WC: Assessment of parent-of-origin effects in linkage analysis of quantitative traits. Am J Hum Genet. 2001, 68: 951-962. 10.1086/319508.
Rowe SJ, Pong-Wong R, Haley CS, Knott SA, De Koning DJ: Detecting dominant QTL with variance component analysis in simulated pedigrees. Genet Res. 2008, 90: 363-374. 10.1017/S0016672308009336.
Bulmer MG: The effect of selection on genetic variability. Am Nat. 1971, 105: 201-10.1086/282718.
Nicholls RD, Knepper JL: Genome organization, function and imprinting in Prader-Willi and Angelman syndromes. Annu Rev Genomics Hum Genet. 2001, 2: 153-175. 10.1146/annurev.genom.2.1.153.
Sewalem A, Morrice DM, Law A, Windsor D, Haley CS, Ikeobi CO, Burt DW, Hocking PM: Mapping of quantitative trait loci for body weight at three, six, and nine weeks of age in a broiler layer cross. Poult Sci. 2002, 81: 1775-1781.
Ikeobi CON, Woolliams JA, Morrice DR, Law A, Windsor D, Burt DW: Quantitative trait loci for meat yield and muscle distribution in a broiler layer cross. Livest Prod Sci. 2004, 87: 143-151. 10.1016/j.livprodsci.2003.09.020.
Wilson HR: Interrelationships of egg size, chick size, posthatching growth and hatchability. Worlds Poult Sci J. 1991, 47: 20-10.1079/WPS19910002.
Pakdel A, Van Arendonk JA, Vereijken AL, Bovenhuis H: Direct and maternal genetic effects for ascites-related traits in broilers. Poult Sci. 2002, 81: 1273-1279.
Hartmann C, Johansson K, Strandberg E, Rydhmer L: Genetic correlations between the maternal genetic effect on chick weight and the direct genetic effects on egg composition traits in a White Leghorn line. Poult Sci. 2003, 82: 1-8.
Jacobsson L, Park HB, Wahlberg P, Fredriksson R, Perez-Enciso M, Siegel PB: Many QTLs with minor additive effects are associated with a large difference in growth between two selection lines in chickens. Genet Res. 2005, 86: 115-125. 10.1017/S0016672305007767.
Abasht B, Pitel F, Lagarrigue S, Le Bihan-Duval E, Le Roy P, Demeure O, Vignoles F, Simon J, Cogburn L, Aggrey S, Vignal A, Douaire M: Fatness QTL on chicken chromosome 5 and interaction with sex. Genet Sel Evol. 2006, 38: 297-311. 10.1051/gse:2006005.
Heuven HC, Bovenhuis H, Janss LL, Van Arendonk JA: Efficiency of population structures for mapping of Mendelian and imprinted quantitative trait loci in outbred pigs using variance component methods. Gen Sel Evol. 2005, 37: 635-655. 10.1051/gse:2005019.
Aerts J, Megens HJ, Veenendaal T, Ovcharenko I, Crooijmans R, Gordon L, Stubbs L, Groenen M: Extent of linkage disequilibrium in chicken. Cytogenet Genome Res. 2007, 117: 338-345. 10.1159/000103196.
This work has made use of the resources provided by the Edinburgh Compute and Data Facility (ECDF). http://www.ecdf.ed.ac.uk/. The ECDF is partially supported by the eDIKT initiative.
The authors would like to thank the reviewers for their helpful comments, Biotechnology and Biological Sciences Research Council, Research Councils UK, and Genesis-Faraday for funding, and Cobb for funding and poultry data.
The authors declare that they have no competing interests.
SJR carried out all of the data analysis, wrote and prepared the manuscript for submission, RP-W wrote the rtools software and aided with data analysis, DJK led the project, wrote the FORTRAN program incorporating the software and evaluated initial rounds of the manuscript. SJR, DJK, RP-W, CSH and SAK were involved in experiment design, critical evaluation and final manuscript revision. All authors read and approved the final manuscript.