| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
ANIMAL GENETICS |
* Departament de Ciència Animal i dels Aliments, Universitat Autònoma de Barcelona, 08193 Bellaterra, Barcelona, Spain;
and
Àrea de Producció Animal, Centre UdLl-IRTA, 25198, Spain; and
and
Centre de Tecnologia de la Carn, IRTA, 17121 Monells, Girona, Spain
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResults and DiscussionImplicationsLiterature Cited |
|---|
Key Words: Pigs • Quantitative Trait Loci • Selected Population
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionImplicationsLiterature Cited |
|---|
). Recently, meat quality has become an important issue in genetic selection due to the demands of consumers. Marker assisted selection (MAS) might be a useful tool to improve selection efficiency for meat quality traits, which are usually difficult to measure. Marker linkage information required for MAS might be acquired through the identification of carcass and meat quality QTL. In recent years, F2 crosses between divergent populations have been used to locate QTL related to growth, fatness (Andersson et al., 1994; Bidanel et al., 2001; Varona et al., 2002), and meat quality traits (Grindflek et al., 2001; de Koning et al., 2001; Clop et al., 2003). These experiments have yielded very valuable information in terms of positioning major QTL that segregate in one or several of these crosses (Walling et al., 2000).
A relevant question that has been recently addressed by several scientific teams is whether QTL that have been characterized in crosses between highly divergent pig breeds are also segregating in highly selected commercial populations. For instance, Evans et al. (2003) analyzed 11 genomic regions for which QTL had been previously reported in 10 Hampshire, Large White, Landrace, Pietrain, and Meishan populations. Their findings confirmed the existence of several of these QTL for growth, backfat, and carcass traits. A second analysis of this experiment with a variance component method yielded similar results, demonstrating the consistency of this approach (de Koning et al., 2003). The main goals of our experiment were to 1) confirm the existence of these previously reported carcass and meat quality QTL in one highly selected Landrace population, and 2) investigate the existence of additional QTL that might be specific to this population.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionImplicationsLiterature Cited |
|---|
Five boars were mated to 71 sows, producing 470 F1 individuals. An average of 94 offspring were obtained from each male. Both males and females were reared in batches according to their age, under normal intensive conditions and fed ad libitum at the experimental farm of Nova Genètica. Males were not castrated and were slaughtered at approximately 179 d of age.
Recording of Phenotypes
Both carcass and meat quality traits were recorded for each individual. Carcass traits were weight at slaughter, carcass weight, carcass length, and backfat thickness (BFT); a Fat-O-Meat’er (SFK Technology, Herlev, Denmark) measurement of the carcass lean percent; fat thickness in the loin area, the cervical region, and in the last rib region; and weight of ham, shoulders, cutlet, ribs, and bacon. Several meat quality traits related to the mechanical and chemical properties of the muscle were also recorded. Chemical composition of the muscle was determined from a sample of the semimembranosus muscle. Samples were homogenized and lyophilized, and fat, protein, OM, and DM were subsequently determined (AOAC, 1990). Mechanical characteristics influencing meat texture were measured with a Texture Analyser TA.TX2 (Stable Micro Systems, Godalming, U.K.) according to the Warner-Bratzler test (Moller, 1981). Samples from the LM were cooked in a water bath at 80°C for 1 h and were subsequently cooled at room temperature. Six 2 cm x 1 cm x 1 cm pieces of each sample were cut in the direction of the muscular fibers, and the following traits were recorded: shear force, which is related to the myofibrillar components of the muscle; maximum shear force, which is related to the proportion of connective tissue; total work required to shear the sample; and shear firmness. Color was measured with a Minolta spectrophotometer at 24 h postmortem: lightness (L), red tendency (a), and yellow tendency (b) were recorded in duplicate (L2, a2, b2) to minimize measurement errors.
Extraction of DNA
Four hundred microliters of blood was washed with 500 µL of TE buffer (10 mM Tris•HCl, pH = 8, and 1 mM EDTA) until a white pellet of cells was obtained. Subsequently, cells were lysed with 40 µg of proteinase K and 400 µL of buffer K (50 mM KCl, 10 mM Tris•HCl, 0.5% Tween 20) for 5 h at 56°C. Genomic DNA was phenol-chloroform extracted and precipitated with 25 µL of 2 M NaCl and 800 µL of ethanol. After 10 min of centrifugation at 15,000 x g, genomic DNA was washed with 70% ethanol and resuspended in 100 µL of TE.
Genotyping
Twenty-three microsatellites were selected according to their chromosomal position and their information content in the analyzed Landrace population. Quantitative trait loci for carcass and meat quality traits had been previously described for the chromosomal regions analyzed in the current work (Table 1). The 546 pigs were genotyped for the complete set of microsatellites and for the Ryr1 gene, according to Fujii et al. (1991). Amplification reactions were carried out in an ABI PRISM 877 Integrated Thermal Cycler (Applied Biosystems, Warrington, U.K.), and were analyzed with the Genescan 3.7 software in a capillary electrophoresis device with fluorescent detection (ABI Prism 3100 genetic analyzer, Applied Biosystems, Warrington, U.K). Genotypes were stored in the Gemma database (Iannuccelli et al., 1996).
|
. We compared two models using a Bayes factor. First, we considered a mixed inheritance model:
 | [1] |
where X, Z1, and Z2 = the incidence matrices, ß = the systematic effects (batch and Ryr1 genotype), u = the polygenic effects, q = the effects associated to a genome segment, and e = the residuals; u ~ N(0, A
2u), A being the polygenic relationship matrix and 2u the polygenic genetic variance and q ~ N(0, Q2q), Q being the relationship matrix associated with the genome segment, 2q the polygenic genetic variance, e ~ N(0, I2e), and 2e is the residual variance. Matrix Q was calculated using the algorithm described by Pérez-Enciso et al. (2000b).
This model (p1) can be reparameterized as y = Xß + e*, where e* = Z1u + Z2q + e and, consequently:
 |
 |
where 
= 2u/2p is the proportion of polygenic variation, and 2p is the phenotypic variance (2u + 2q + 2e
Records and parameters are jointly distributed as:
 |
where
 |
 |
 |
Note that the parametric space for both heritabilities (h2 q, h2u) is a triangle, and the flat density to assure a volume equal to one must be two. Note also that, assuming prior independence, marginal priors of h2q and h2u are:
 |
 |
The second model (p2) that we have assumed is:
 | [2] |
which can be reduced to y = Xß + e*, where e* = Z1u+ e and, consequently:
 |
 |
 |
where priors for ß and 2p are the same as in Model [1]. Prior distribution for h2u is:
 |
where U denotes a uniform distribution. It should be noted that Model [2] is a special case of Model [1] when h2q = 0.
According to Varona et al. (2001), the Bayes factor of Model [1] against Model [2] is:
 |
because p1(h2q = 0) = 2.
As previously reported by Varona et al. (2001), only the analysis with the complex model (Model [1]) is required. In this situation, only the Bayesian calculations with the complex model are needed. The calculation of the posterior distribution p1(h2q/y) was performed using a Gibbs sampler (Gelfand and Smith, 1990) with a Metropolis-Hasting step (Hastings, 1970) used to sample from the conditional distribution of the heritabilities (h2qh2u). A total of 25,000 iterations were performed after discarding the first 5,000. All correlated samples were used to calculate the posterior distributions using the ergodic property of the chain (Gilks et al., 1996). Convergence was checked using the algorithm of Raftery and Lewis (1992).
The results of Minolta color components (L, L2, a, a2, b, and b2) were calculated using the meta-analysis approach described by Varona et al. (2001). This approach assumed that the data are two identical replicates, for this reason the joint Bayes factor is the multiplication of the Bayes factor obtained for each replicate.
The posterior probability is directly related to the Bayes factor if we assume that prior probabilities of both models are 0.5. The posterior probability (PP) of the QTL model is calculated as follows:
 |
As a result, we have considered as relevant Bayes factors higher than 3, where the PP of the QTL model is >0.75. Bayes factors from 3 to 20 (PP from 0.75 to 0.95) can be considered as suggestive of linkage, those from 20 to 150 (PP from 0.95 to 0.99) indicate linkage, and if the Bayes factor is higher than 150 (PP >0.99), there is a very strong evidence of QTL linkage. A Bayes factor <1.0 is associated with posterior probability smaller than 0.5, and it is interpreted as absence of a QTL in the analyzed region.
|
Results and Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionImplicationsLiterature Cited |
|---|
), especially with the availability of genetic expression data from microarrays. There is a straightforward alternative, which is equivalent to the Bonferroni test. This method considers the P(Ho) in all cases to be equal to 0.5 (that is, the probability of no QTL in all the locations and all the traits is equal to 0.5). If we make this assumption in our QTL experiment, the prior odds for each of the 280 comparisons should be 5.15 x 10–85, making it almost impossible to obtain posterior odds greater than one. Nevertheless, this prior assumption would involve assuming that there are no genes affecting any of the traits and in any of the locations with probability 0.5, a hypothesis that seems unrealistic. Moreover, it would imply that all of the traits are uncorrelated, and this is clearly false in all of the QTL experiments. Less conservative prior odds can be defined, but this is not straightforward. As a conservative rule, we consider a BF greater than 3 to be a confident indicator of a QTL. Strictly, a BF greater than 1, implies that the probability of the QTL model is greater than the no-QTL model, but we do not consider this as suggestive of linkage. The PP of the no-QTL model is below 0.25 when the BF is greater than 3. This PP of the no-QTL model is related with the false discovery rate (FDR).
Genome Scan Analysis
We have performed a QTL search for 24 carcass and meat quality traits in a commercial Landrace pig population (Table 2). Ten genomic regions (pig Chromosomes 1, 2, 3, 4, 6, 7, 8, 9, 10, 13), for which growth, fatness, and meat quality QTL had been previously reported (Table 1
), were analyzed in this experiment.
|
; Jeon et al., 1999; Milan et al., 2002). The identification of the causal mutation of the IGF2 QTL, an intronic polymorphism with influence on muscle mass recently described by van Laere et al. (2003), will be fundamental for establishing if it is the same one that is segregating in our population.
We did not find any QTL related to fatness in the SW2443-IGF2 interval, in spite of the fact that other authors reported QTL affecting this trait in a location distal to the IGF2 region (de Koning et al., 1999; Rattink et al., 2000). This discrepancy might be attributed to the fact that we are not analyzing the same chromosomal region or that this QTL is fixed in our population.
With regard to meat quality traits, we found QTL for shear firmness (PP > 0.99). To the best of our knowledge, there are no reports describing shear firmness or shear force QTL on Chromosome 2. Because we are unable to estimate precisely the location of this meat tenderness QTL, we do not know whether Chromosome 2 contains one single pleiotropic QTL regulating muscular mass and meat tenderness or several neighboring QTL influencing these traits.
Chromosome 7.
Quantitative trait loci with influence on fatness (measured in the last rib region, PP > 0.75) and color (b, PP > 0.75) have been detected in our commercial population. A QTL located in this region and affecting fatness and diverse growth traits, has been consistently described by several authors (Rohrer and Keele, 1998a; Bidanel et al., 2001; Rattink et al., 2000). Rothschild et al. (1995) and Wang et al. (1998) also reported QTL for fatness and color in this region. Moreover, Óvilo et al. (2002) reported a QTL with effect on color on Chromosome 7. Because lipid oxidation might result in a lighter meat color, we might hypothesize the existence of one pleiotropic QTL regulating lipid content and meat color, although we cannot completely rule out the existence of two different QTL coexisting in the same genomic region.
Chromosome 8.
We found suggestive evidence for the existence of QTL influencing BFT and weight of ribs (PP > 0.75). Quantitative trait loci with effects on body proportion (Andersson-Eklund et al., 1998), growth and weight (Quintanilla et al., 2002), growth and fatness (Bidanel et al., 2001), and fatness (Rohrer and Keele, 1998a) had been previously described in this region. However, there is a lack of consistency in the precise location of these QTL, a feature that suggests the existence of several QTL affecting growth and fatness with a differential pattern of segregation in each one of the analyzed populations. As a result, although we have found associations between this chromosomal region and growth and fatness, it is difficult to find a correspondence between these QTL and the ones reported in the scientific literature.
Chromosome 9.
Suggestive QTL for BFT (PP > 0.75) was found. Quantitative trait loci for fat deposition have been described in this region (Rohrer and Keele, 1998a) and might correspond to the one we have detected. In fact, Evans et al. (2003) detected QTL for growth and carcass performance at this region, in close agreement with our results.
Chromosome 10.
We have found suggestive QTL for BFT (PP > 0.75). Rohrer and Keele (1998a) reported a QTL for BFT in a distal position, which might correspond to the one we have found. Moreover, Evans et al. (2003) found one QTL for carcass length; however, the low Bayes factor associated with this QTL makes it difficult to establish correspondence with previous experiments.
Maintenance of Genetic Diversity for Carcass and Meat Quality QTL in a Commercial Population Selected for Growth and Leanness
Although a large number of QTL related to traits of economic interest have been found in pigs by using crosses between divergent populations, the segregation of these QTL in the highly selected populations used in commercial livestock breeding has not been analyzed until recently (Evans et al., 2003; Nagamine et al., 2003). In the current work, we have analyzed growth and fatness traits, and an extensive repertoire of meat quality traits mainly related to meat color, chemical composition, and texture, that have not been measured in previous QTL experiments with commercial populations.
Our results confirm the existence of at least one QTL that had been previously described in divergent crosses in a Landrace population, the QTL for leanness on Chromosome 2. Moreover, some of the other suggestive effects, with PP between 0.75 and 0.95, might correspond to previously described QTL (e.g., QTL for color on Chromosome 7). In contrast, in some cases, it is possible that we have found new QTL that had not been detected in previous studies (e.g., QTL for weight on Chromosome 13).
These results are in good agreement with similar studies undertaken by Nagamine et al. (2003) and Evans et al. (2003). Nagamine et al. (2003) studied the same regions on Chromosomes 4 and 7 that we analyzed in our population. They analyzed five different commercial populations and found evidence of segregating QTL on Chromosomes 4 and 7 within two and five populations, respectively. Moreover, Evans et al. (2003) analyzed the same chromosomal regions tested in our experiment in 10 selected populations from diverse breeds. They found clear evidence indicating the existence of many growth and fatness QTL in these purebred populations. Quantitative trait loci affecting pH and electric conductivity were also detected, demonstrating for the first time the existence of meat quality QTL in commercial populations.
The main conclusion that can be drawn from the QTL analyses carried out by us and others is that commercial populations still retain an important amount of genetic variation for traits for which they have been selected for a considerable period of time. This finding might be explained by the fact that pure-bred populations are usually selected according to diverse criteria, a feature that diminishes the probability of allele fixation for the QTL that influence a single trait. Furthermore, selection criteria have changed through time, and introgression of foreign material in supposedly "pure" breeds is the rule rather than the exception. The existence of pleiotropic genes with alleles that are favorably correlated to some traits and unfavorably with others might also explain the maintenance of genetic diversity in selected pig populations.
The number and magnitude of the QTL we identified in this Landrace purebred population were lower than those reported in divergent crosses. This result is not surprising because divergent crosses involve breeds that have largely diverged at the phenotypic and genetic levels. Obviously, most of these divergent crosses are not usually performed by pig breeders and, as a consequence, the QTL information they provide might have a limited practical application in the pig industry. Our data show that part of this information generated in divergent crosses can be extrapolated to commercial populations. For instance, the Chromosome 2 muscular mass QTL reported in European Wild Boar x Large White and Large White x Pietrain intercrosses (Jeon et al., 1999; Nezer et al., 1999) segregates in the Landrace population analyzed in our experiment. The existence of particular QTL alleles in purebred commercial populations might be influenced by a plethora of factors, such as their demographic history, geographical location, and breeding goals.
Interestingly, we found that growth and fatness QTL are much more important and abundant than meat quality QTL in our purebred population (Table 2). This result, which is in close agreement with other QTL data reported in divergent crosses (Andersson-Eklund et al., 1998; Óvilo et al., 2002; Varona et al., 2002), might be interpreted in the sense that meat quality traits have a different genetic architecture than growth and fatness traits. In fact, the average heritability values are usually higher for body composition traits, such as backfat thickness (h2 = 0.41 to 0.45) and carcass length (h2 = 0.56 to 0.57), than for meat quality traits such as pH (h2 = 0.21), tenderness (h2 = 0.26), and color (h2 = 0.28; Sellier, 1998).
In the future, the understanding of the genetic architecture of quantitative traits will probably evolve to identification of the genetic polymorphisms that influence them and the elucidation of the transcriptional and post-transcriptional mechanisms that regulate their expression.
|
Implications |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionImplicationsLiterature Cited |
|---|
|
Footnotes |
|---|
2 Correspondence—phone: 34-93-5812087; fax: 34-93-5812106; e-mail: oriol.vidal{at}uab.es.
Received for publication November 25, 2003. Accepted for publication October 19, 2004.
|
Literature Cited |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionImplicationsLiterature Cited |
|---|
This article has been cited by other articles:
|
|
 |
|
  Y. Uemoto, Y. Nagamine, E. Kobayashi, S. Sato, T. Tayama, Y. Suda, T. Shibata, and K. Suzuki Quantitative trait loci analysis on Sus scrofa chromosome 7 for meat production, meat quality, and carcass traits within a Duroc purebred population J Anim Sci, November 1, 2008; 86(11): 2833 - 2839. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Gallardo, R. N. Pena, M. Amills, L. Varona, O. Ramirez, J. Reixach, I. Diaz, J. Tibau, J. Soler, J. M. Prat-Cuffi, et al. Mapping of quantitative trait loci for cholesterol, LDL, HDL, and triglyceride serum concentrations in pigs Physiol Genomics, November 1, 2008; 35(3): 199 - 209. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Estany, M. Tor, D. Villalba, L. Bosch, D. Gallardo, N. Jimenez, L. Altet, J. L. Noguera, J. Reixach, M. Amills, et al. Association of CA repeat polymorphism at intron 1 of insulin-like growth factor (IGF-I) gene with circulating IGF-I concentration, growth, and fatness in swine Physiol Genomics, October 19, 2007; 31(2): 236 - 243. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. A. Kuehn, G. A. Rohrer, D. J. Nonneman, R. M. Thallman, and K. A. Leymaster Detection of single nucleotide polymorphisms associated with ultrasonic backfat depth in a segregating Meishan x White Composite population J Anim Sci, May 1, 2007; 85(5): 1111 - 1119. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. J. van Wijk, B. Dibbits, E. E. Baron, A. D. Brings, B. Harlizius, M. A. M. Groenen, E. F. Knol, and H. Bovenhuis Identification of quantitative trait loci for carcass composition and pork quality traits in a commercial finishing cross J Anim Sci, April 1, 2006; 84(4): 789 - 799. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. J. Bequette, N. E. Sunny, S. W. El-Kadi, and S. L. Owens Application of stable isotopes and mass isotopomer distribution analysis to the study of intermediary metabolism of nutrients J Anim Sci, April 1, 2006; 84(13_suppl): E50 - E. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
