SEARCH

SEARCH BY CITATION

Keywords:

  • horse;
  • microsatellite;
  • recurrent exertional rhabdomyolysis;
  • Thoroughbred;
  • tying-up syndrome

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results and discussion
  6. Conclusion
  7. Acknowledgements
  8. Conflicts of interest
  9. References

Tying-up syndrome, also known as recurrent exertional rhabdomyolysis in Thoroughbreds, is a common muscle disorder for racehorses. In this study, we performed a multipoint linkage analysis using LOKI based on the Bayesian Markov chain Monte Carlo method using 5 half-sib families (51 affected and 277 nonaffected horses in total), and a genome-wide association study (GWAS) using microsatellites (144 affected and 144 nonaffected horses) to map candidate regions for tying-up syndrome in Japanese Thoroughbreds. The linkage analysis identified one strong L-score (82.45) between the loci UCDEQ411 and COR058 (24.9–27.9 Mb) on ECA12. The GWAS identified two suggestive genomic regions on ECA12 (24.9–27.8 Mb) and ECA20 (29.3–33.5 Mb). Based on both results, the genomic region between UCDEQ411 and TKY499 (24.9–27.8 Mb) on ECA12 was the most significant and was considered as a candidate region for tying-up syndrome in Japanese Thoroughbreds.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results and discussion
  6. Conclusion
  7. Acknowledgements
  8. Conflicts of interest
  9. References

Tying-up syndrome is a common intermittent condition that primarily affects the muscles in horses. It is characterized by clinical signs ranging from slight stiffness to immobility, signs of pain and reluctance to move, which may develop after mild-to-moderate exercises (McLean 1973; Rossdale et al. 1985; Harris 1989, 1991). It is also known as azoturia, exertional rhabdomyolysis, paralytic myoglobinuria, Monday morning disease and exertional myopathy, because the clinical signs of this syndrome vary. Tying-up syndrome occurs because of specific inherited abnormalities and can be classified into two distinct types: polysaccharide storage myopathy (PSSM) in Quarter horses and recurrent exertional rhabdomyolysis (RER) in Thoroughbreds. A mutation in the gene encoding skeletal muscle glycogen synthase (GYS1) was recently reported to be strongly associated with PSSM in Quarter horses and Belgian Draught horses (McCue et al. 2008a,b; Stanley et al. 2009). This GYS1 point mutation appears to be a gain-of-function mutation that results in the accumulation of a glycogen-like, less-branched polysaccharide in skeletal muscles. It is inherited as a dominant trait (McCue et al. 2008b).

For more than a century, tying-up syndrome, also known as RER, in Thoroughbreds has been recognized as a syndrome of muscle pain and cramping associated with exercises (McKenzie et al. 2003). This syndrome occurs in approximately 5.0% of all Thoroughbreds (MacLeay et al. 1999; McGowan et al. 2002) in the United Kingdom and United States of America and is hence responsible for substantial economic loss. A statistical genetic analysis of pedigrees indicated that the susceptibility of Thoroughbreds to tying-up syndrome is a heritable trait (Oki et al. 2005). Although the occurrence of this syndrome may be influenced by multiple factors such as sex, temperament and diet, an in vitro muscle contracture test involving various breeding trials suggested that tying-up syndrome in Thoroughbreds can be modelled as a genetic trait with an autosomal dominant mode of inheritance (Dranchak et al. 2005). Many studies based on proteomic approaches have been performed to identify the susceptibility genes for tying-up syndrome (Lentz et al. 2002; McKenzie et al. 2002, 2004). However, no critical susceptibility genes have yet been identified. A recent linkage analysis using the microsatellites around the RYR1, CACNA1S and ATP2A1 genes, which were expected to be susceptibility genes because of their role in myoplasmic calcium regulation in skeletal muscles, did not reveal associations with tying-up syndrome in Thoroughbreds (Dranchak et al. 2006).

Many significant advances have recently been achieved in horse genome mapping, and linkage maps of horses have been created (Penedo et al. 2005; Swinburne et al. 2006). Horse–human comparative maps were published in 2007 and 2008 (Tozaki et al. 2007a; Raudsepp et al. 2008). With the complete horse genome sequence available to the equine genome community, we expect an acceleration in discoveries of genetic traits of interest to horse breeders (Wade et al. 2009). In our previous studies, we proposed a genome-wide association study (GWAS) that can be systematically performed by combining primary screening for genome-wide markers with pooled DNA samples and secondary screening for candidate markers implicated from the genome-wide scan in Thoroughbreds (Tozaki et al. 2005, 2007b).

In this study, we aimed at identifying candidate genomic regions susceptible for tying-up syndrome in Japanese Thoroughbred populations by performing a Markov chain Monte Carlo (MCMC) linkage analysis and a GWAS using microsatellites.

Materials and methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results and discussion
  6. Conclusion
  7. Acknowledgements
  8. Conflicts of interest
  9. References

Thoroughbreds and DNA purification

Thoroughbreds with and without tying-up syndrome were obtained from the Ritto and Miho Training centers of the Japan Racing Association (JRA). In terms of clinical infrastructure and expertise, these training centers of the JRA offer state-of-the-art horse clinic facilities for the diagnosis and treatment of various equine disorders. All the clinical information regarding the JRA’s racehorses is stored in a computerized database system (JARIS: Japan Racing Information System). Of the 3927 Thoroughbreds registered as JRA racehorses on October 2002, 223 Thoroughbreds with tying-up syndrome were selected for DNA isolation. On the basis of the information available on the relatives of the 223 horses, five paternal half-sib Thoroughbred families (designated A through E) were then constructed for linkage analysis using LOKI (Table 1). The half-sib families were unrelated for at least three generations of paternal pedigree. Furthermore, from these 223 horses, we selected 144 individuals that had no paternal, maternal or sibling relationships with each other for the GWAS. The control population comprised 144 normal horses without tying-up syndrome that had been randomly selected from both training centres; these horses had more racing experience and also had no paternal, maternal or sibling relationships with each other. No statistically significant differences were observed between both populations for the ratio of gender and/or for training centre.

Table 1.   Number of tying-up and non-tying-up offspring for each sire family.
FamilyTying-upNon-tying-upDisease rate (%)
  1. 1These were sampled from 233 affected horses.

A207221.7
B118611.3
C96612.0
D72323.3
E43011.8
Total51127718.4

All blood samples for DNA isolation were stored at −40 °C. Genomic DNA was extracted from the stored blood by using the MagExtractor System MFX-2000 (Toyobo, Osaka, Japan). The concentrations of the DNA samples were measured twice using a multi-wavelength (200–310 nm) absorption spectrometer (DU®7500 Spectrophotometer; Beckman Coulter, Brea, CA, USA) and diluted to a concentration of 25 ng/μl. The concentrations of the samples were measured again before they were combined for the case–control association studies. Finally, the DNA samples were combined with their respective pools. The concentration of all the pooled DNA samples was adjusted to 25 ng/μl.

Microsatellites

For the linkage analysis, 117 microsatellites on autosomal chromosomes were selected from the workshop linkage map (Penedo et al. 2005). For the GWAS performed using pooled DNA, 986 microsatellites on autosomal chromosomes were used as genetic markers for the first screening for the genome-wide scan. The markers for the GWAS were mainly selected from the workshop linkage map, with an average interval of 6.3 cM between markers using 766 microsatellites. The microsatellites mapped on the AHT linkage map (Swinburne et al. 2006) and a human–horse comparative map (Tozaki et al. 2007a) were also added to increase the marker density. The added marker improved marker intervals of over 10 cM in the workshop linkage map, and it was expected that all the marker intervals would become less than 10 cM in the first screening. In the second screening, the additional microsatellites on the chromosomes that showed statistical significance for tying-up syndrome in the first screening were used as markers for individual genotyping.

Microsatellite genotyping

For microsatellite genotyping, we used the following 3 primers: a sequence-specific forward primer conjugated with a 5′-TGACCGGCAGCAAAATTG-3′ tail at its 5′ end, a sequence-specific reverse primer, and a FAM-labelled 5′-TGACCGGCAGCAAAATTG-3′ primer (Applied Biosystems, Foster City, CA, USA). The underlined sequences were used for fluorescence detection. PCR was performed as described in Tozaki et al. (2001). The reaction products were analysed using the ABI 3130 genetic analyzer (Applied Biosystems), and the genotyping data were processed by the Genotyper software (Applied Biosystems) to identify alleles and to measure the peak height of the alleles for each microsatellite.

Linkage analysis using LOKI

We carried out multipoint linkage analysis assuming multiple putative quantitative trait loci (QTL) simultaneously in the oligogenic model. We used the LOKI 2.4 program (http://www.stat.washington.edu/thompson/Genepi/Loki.shtml), which is based on the Bayesian reversible jump MCMC method (Heath 1997), and can estimate the posterior distribution of a number of parameters of interest (e.g. the number of QTL contributing to the trait, their location and the genotype effects for each QTL). The analytical method basically followed the study of Shmulewitz et al. (2006). The LOKI results are reported as L-scores, which, as estimates of the Bayes factor (Kass & Raftery 1995), show the posterior ratio between the probability that a QTL signal is ‘real’ and that it is because of chance alone. To compute the L-scores, the prior probability P of finding a QTL linked to a 1-cM bin is 1/t, where t is the total map length of the genome. For a given iteration, the prior probability that at least 1 QTL is linked to a 1-cM bin is 1 − (1 − 1/t)n, where n is the number of QTL in the model at that iteration. The posterior probability q is 1 or 0, depending on whether or not a QTL is linked. The posterior (p)/prior (q) ratio for each linkage group is then averaged over iterations to give the L-score (for more information, see Neuman et al. (2002) or the LOKI usage manual). For all the other parameters, uniform priors were adopted. According to the Bayes factor calibration tables, an L-score of >10 is suggestive of linkage while an L-score of >20 provides evidence of linkage (Raftery 1996).

In this multipoint linkage analysis, phenotypic information regarding tying-up syndrome was recorded as binary data (0.0 or 1.0) but handled as continuous quantitative characters during the computation by the LOKI program. The pedigrees of all the Thoroughbreds were traced back for three generations. The total number of MCMC chain lengths was 30 000, with a burn-in of 10 000. The convergence of the estimates was confirmed by at least 2 runs with different starting seeds, especially for the QTL location with the largest L-score. The linkage analysis was separately performed for each half-sib family.

GWAS

The differences in the total allele content (TAC) in the microsatellites between different DNA pools were evaluated by calculating ΔTAC (Collins et al. 2000; Tozaki et al. 2007b). The peak height of each allele in a pool was determined from an electrophoretogram, and the sum of all the peak heights in the pool was calculated. The height of each allele was then divided by the sum of all peak heights and multiplied by 100 to obtain the allele percentage. The two pools were then compared, and the absolute value of the difference in the percentage for the two pools was computed for each allele. ΔTAC was then computed using the following formula:

  • image

where X1 is the allele percentage in pool 1 and X2 is the allele percentage in pool 2. In the first screening, a ΔTAC value of >10 was regarded as positive (Collins et al. 2000; Tozaki et al. 2007b). This procedure was repeated for each of the markers.

A chi-square test of association using the genotypes of individual horses was performed to compare the allele frequency between the case and control populations for seven microsatellites that were chosen during the first screening. Additionally, 13 microsatellites located near the positive microsatellites in the first screening were included. The SNPAlyze ver. 7.0 Standard program (Dynacom, Yokohama, Japan) was used for computation.

In addition, the permutation test was performed to test for deviation in the allele frequencies of markers and haplotypes. Haplotype frequencies for two markers were estimated by the maximum-likelihood method with an expectation-maximization algorithm. Distribution of the test statistics was estimated by evaluating the statistics for a random sampling of 10 000 iterated permutations; the total numbers of both cases and controls were fixed to avoid false-positive results in multiple testing. We calculated the LD between pairs of markers by using the standard definition of D’. The SNPAlyze ver. 7.0 Standard program was used for computation.

Results and discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results and discussion
  6. Conclusion
  7. Acknowledgements
  8. Conflicts of interest
  9. References

Linkage analysis with LOKI

The MCMC linkage analysis performed using LOKI identified a significant L-score on ECA12 (82.45) only in family B (Table 2). The location for the best L-score was 26.5 cM on ECA12, which produced a sharp peak (Fig. 1) between the loci UCDEQ411 and COR058. This suggests that the region between these markers (24.9–27.9 Mb) on ECA12 is a candidate region for tying-up syndrome in Japanese Thoroughbreds. In the other families, namely, families A, C, D and E, a suggestive L-score (≥10) for linkage is not observed at this position (Table 2). The expected number of QTL for each family is listed in Table 3. On average, one or two QTL are expected to segregate in these families. However, considering the low L-scores in the families A, C, D and E, the expected number of QTL in those families would not be suitable for estimation. Therefore, the expected number of QTL was calculated as 1.51 for family B, suggesting the presence of at least one QTL segregating in family B.

Table 2.   Maximum L-scores and their map position on each chromosome by linkage analysis using LOKI.
ECAFamily AFamily BFamily CFamily DFamily E
L-scorePosition (cM)L-scorePosition (cM)L-scorePosition (cM)L-scorePosition (cM)L-scorePosition (cM)
 15.1925.51.86145.51.4430.54.9764.51.96132.5
 21.6321.51.8578.51.554.51.4255.51.8244.5
 38.1528.51.2264.51.8014.53.0884.51.73100.5
 47.5038.52.1320.51.746.56.4012.52.0695.5
 52.2328.50.9188.56.5295.51.840.51.7344.5
 62.5070.53.9597.52.2389.51.913.51.5482.5
 71.1928.51.3540.51.4923.52.8740.51.5812.5
 84.7850.50.8221.51.6088.51.7885.51.77106.5
 90.7210.51.1011.51.9355.51.207.52.368.5
102.920.50.6523.51.3723.55.7023.51.814.5
111.625.50.7816.51.376.51.177.52.265.5
122.3316.582.4526.51.6210.51.2941.51.2814.5
131.149.50.347.51.0717.53.9621.51.610.5
141.9868.51.0290.51.7577.51.32124.51.8825.5
152.611.51.0143.52.7476.51.9417.51.4137.5
163.0952.52.4242.52.7985.53.0873.51.8281.5
171.9813.50.7422.51.3927.52.0538.52.0829.5
182.004.57.8940.51.428.54.9622.51.804.5
191.681.51.2315.51.386.51.632.51.9017.5
202.132.50.995.51.6949.51.1468.51.2151.5
212.746.50.9666.53.8163.52.2430.51.8816.5
222.3344.51.1760.51.344.51.9663.51.5251.5
231.940.50.260.51.000.50.580.51.780.5
241.122.50.5710.51.617.50.6234.50.8919.5
250.9821.53.2330.51.2427.51.0929.52.0629.5
262.018.51.652.51.151.51.528.51.0911.5
271.9414.51.246.51.222.52.974.51.4411.5
280.0114.50.016.50.012.50.014.50.0111.5
291.5719.51.0615.51.3148.51.697.51.7937.5
301.272.50.5911.51.6719.52.0222.51.748.5
310.8710.51.452.51.3238.51.3838.50.8820.5
image

Figure 1.  L-score for ECA12 in Family B constructed by MCMC linkage analysis.

Download figure to PowerPoint

Table 3.   The expected number of QTL for tying-up syndrome with LOKI analyses.
QTL Number0123456Average
Family A0.00000.61590.27570.08000.02230.00500.00101.5275
Family B0.00000.62030.27850.07250.02040.00460.00311.5180
Family C0.01250.34420.29580.18540.08390.04240.02002.1596
Family D0.00000.52080.29260.11810.04650.01350.00321.7330
Family E0.00670.35950.29370.17900.09620.03840.01672.1609

If only outbred sire half-sib family populations are used in the linkage analysis, then only the sire heterozygous for the target QTL will be useful for QTL detection. In families A, C, D and E, the QTL of the sires used could be homozygous on ECA12. In addition, considering the small sample size for each family (Table 1), the power for the detection of QTL would be low, especially for families D and E.

GWAS

In the first screening, pooled DNA samples from the case and control populations were used for the GWAS involving 986 microsatellites. From these, seven microsatellites with a ΔTAC of >10, namely, COR028 on ECA3; TKY331 on ECA5; TKY499, COR058 and UCDEQ411 on ECA12; UM011 on ECA20; and L12.2 on ECA29, were subjected to the second screening with individual genotyping of the 144 case and control samples. The microsatellites UM011 (= 0.0055) on ECA20 and TKY499 (P = 0.0258) on ECA12 showed statistically significant associations (< 0.05) for tying-up syndrome without Bonferroni correction; however, UM011 showed a statistically significant association only with Bonferroni correction (i.e. significant level with 0.05/7) in the second screening. Those suggestive genomic regions were then subjected to the third screening. The third screening was a permutation test to evaluate the differences in the haplotype frequencies of two neighbouring microsatellites (Table 4). Several microsatellites located near the positive markers on each chromosome, namely, AHT017 and UCDEQ497 on ECA12; and LEX064, TKY136, UMNe056, TKY694, HMS082, UMNe065, TKY1115, TKY821, TKY539, TKY547 and TKY1048 on ECA20, were also genotyped (Table 4). Among these microsatellites, the microsatellites located in the region between 29.3 and 34.9 Mb on ECA20 also showed a statistically significant association (< 0.05) for tying-up syndrome. For the microsatellite pairs, significant permutation P values were also observed on ECA12 and ECA20 (Table 4), i.e. UCDEQ411/TKY499 (= 0.0051) on ECA12 and HMS082/UMNe065 (= 0.0169) and UM011/TKY1115 (= 0.0152) on ECA20. The results for UCDEQ411/TKY499 (= 0.0051) on ECA12 corresponded well with the results for linkage analysis, with a high L-score (Fig. 1). Although the linkage analysis did not detect any QTL on ECA20, the results of GWAS showed statistically significant associations between markers on ECA12 and ECA20, suggesting that ECA12 and ECA20 are candidate susceptibility regions for tying-up syndrome in Japanese Thoroughbreds.

Table 4.   Microsatellites tested on chromosome 12 and 20 for tying-up susceptibility.
ECAMarkersMb1cM2cM3P-valueHWEMarker pairsP-value4LD
Tying-upNon-tying-up
  1. 1The genomic position of chromosome 12 and 20 (EquCab2.0) is given in Mbp.

  2. 2The positions of markers along equine chromosome 12 and 20, according to the AHT linkage map, are given in cM.

  3. 3The positions of markers along equine chromosome 12 and 20, according to the international equine gene mapping workshop report, are given in cM.

  4. 4P-values based on permutation test.

12UCDEQ41124.925.637.60.20518.54E-011.24E-01UCDEQ411/TKY4990.00510.25
12TKY49927.8  0.02589.82E-014.17E-01TKY499/COR0580.04290.83
12COR05827.925.646.80.14707.30E-021.25E-02COR058/AHT0170.26890.25
12AHT017  66.40.57789.49E-012.73E-01AHT017/UCDEQ4970.38640.51
12UCDEQ49732.558.070.20.73024.51E-013.42E-01   
20LEX06415.3 49.90.28606.85E-018.26E-01LEX064/TKY1360.79580.22
20TKY136 20.4 0.28359.13E-015.97E-01TKY136/UMNe0560.03930.82
20UMNe05629.323.4 0.02006.81E-017.00E-01UMNe056/TKY6940.08170.74
20TKY69431.2  0.01095.14E-013.43E-01TKY694/HMS0820.06110.90
20HMS082 26.1 0.01905.12E-013.20E-01HMS082/UMNe0650.01690.96
20UMNe06531.5  0.00953.36E-013.22E-01UMNe065/UM0110.17120.61
20UM01133.527.6 0.00556.61E-012.36E-02UM011/TKY11150.01520.78
20TKY111534.9  0.04578.56E-018.85E-02TKY1115/TKY8210.16120.34
20TKY82139.7  0.56756.63E-013.39E-01TKY821/TKY5390.95830.10
20TKY53940.3  0.73756.13E-013.63E-01TKY539/TKY5470.27740.21
20TKY54740.8  0.59209.97E-017.61E-02TKY547/TKY10480.70540.62
20TKY104848.9  0.43408.75E-012.24E-01   

The power of whole-genome LD screening depends on the number of markers analysed, the number of samples in the case and control populations, the extent of LD in the horse genome and the marker density required for effective association. The number of case and control samples used in this study (N = 144) may be rather small to detect the susceptibility loci for tying-up syndrome. Furthermore, we could not clarify the mode of inheritance of the candidate QTL, probably because of the use of the less informative binary phenotype. Further investigations into the Thoroughbreds diagnosed with tying-up syndrome in other racing organizations and countries are expected.

It is known that the linkage analysis is not influenced by population stratification. Therefore, linkage analysis using LOKI is useful for the detection of QTL for complex traits in outbred populations. The combination of linkage analysis with LOKI and our GWAS approach with microsatellites (Tozaki et al. 2007b) can accelerate the detection of QTL or candidate regions for complex traits.

A previous study reported that tying-up syndrome occurred in 7.7% of the JRA’s Thoroughbreds and that the prevalence of this syndrome in fillies, colts and geldings was 11.0%, 5.7%, and 5.8%, respectively (Oki et al. 2005). The gender-based difference in prevalence may support the theory that this syndrome is a sex-linked disorder. Further investigation involving the sex chromosomes will thus be worthwhile. Oki et al. (2005) also clarified that tying-up syndrome is heritable and estimated the heritability to be approximately 0.42 by using the threshold model. It is thought that the ECA12 and ECA20 candidate regions identified in this study may partly explain the heritability of typing-up syndrome.

Conclusion

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results and discussion
  6. Conclusion
  7. Acknowledgements
  8. Conflicts of interest
  9. References

In this study, we performed MCMC linkage analysis and GWAS and identified two candidate regions on horse chromosomes ECA12 and ECA20 for tying-up syndrome in Japanese Thoroughbreds. The results obtained by both methods provided strong evidence for the presence of QTL on ECA12.

Although the recent advances in the use of single-nucleotide polymorphisms (SNPs) in genomic studies are impressive, we still believe that our GWAS approach using microsatellites will be quite useful in cases where sufficient genomic information or tools (such as SNP chips) are not available. The GWAS described here was performed using pooled DNAs and is a cost-effective method that can also be used for many livestock animals.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results and discussion
  6. Conclusion
  7. Acknowledgements
  8. Conflicts of interest
  9. References

We thank the Ritto and Miho Training centers of the JRA for providing samples from their horses as study materials. We also thank the Equine Department of JRA for approving and supporting the study with a grant-in-aid (2005–2007).

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results and discussion
  6. Conclusion
  7. Acknowledgements
  8. Conflicts of interest
  9. References
  • Collins H.E., Li H., Inda S.E., Anderson J., Laiho K., Tuomilehto J. & Seldin M.F. (2000) A simple and accurate method for determination of microsatellite total allele content differences between DNA pools. Human Genetics 106, 21826.
  • Dranchak P.K., Valberg S.J., Onan G.W., Gallant E.M., MacLeay J.M., McKenzie E.C., De La Corte F.D., Ekenstedt K. & Mickelson J.R. (2005) Inheritance of recurrent exertional rhabdomyolysis in thoroughbreds. Journal of the American Veterinary Medical Association 227, 7627.
  • Dranchak P.K., Valberg S.J., Onan G.W., Gallant E.M., Binns M.M., Swinburne J.E. & Mickelson J.R. (2006) Exclusion of linkage of the RYR1, CACNA1S, and ATP2A1 genes to recurrent exertional rhabdomyolysis in Thoroughbreds. American Journal of Veterinary Research 67, 1395400.
  • Harris P. (1989) Equine rhabdomyolysis syndrome. In Practice 11, 38.
  • Harris P.A. (1991) The equine rhabdomyolysis syndrome in the United Kingdom: epidemiological and clinical descriptive information. British Veterinary Journal 147, 37384.
  • Heath S.C. (1997) Markov chain Monte Carlo segregation and linkage analysis for oligogenic models. American Journal of Human Genetics 61, 74860.
  • Kass R.E. & Raftery A.E. (1995) Bayes factors. Journal of the American Statistical Association 90, 77395.
  • Lentz L.R., Valberg S.J., Herold L.V., Onan G.W., Mickelson J.R. & Gallant E.M. (2002) Myoplasmic calcium regulation in myotubes from horses with recurrent exertional rhabdomyolysis. American Journal of Veterinary Research 63, 172431.
  • MacLeay J.M., Sorum S.A., Valberg S.J., Marsh W.E. & Sorum M.D. (1999) Epidemiologic analysis of factors influencing exertional rhabdomyolysis in thoroughbreds. American Journal of Veterinary Research 60, 15626.
  • McCue M.E., Valberg S.J., Miller M.B., Wade C., DiMauro S., Akman H.O. & Mickelson J.R. (2008a) Glycogen synthase (GYS1) mutation causes a novel skeletal muscle glycogenosis. Genomics 91, 45866.
  • McCue M.E., Valberg S.J., Lucio L. & Mickelson J.R. (2008b) Glycogen synthase 1 (GYS1) mutation in diverse breeds with polysaccharide storage myopathy. Journal of Veterinary Internal Medicine 22, 122833.
  • McGowan C.M., Fordham T. & Christley R.M. (2002) Incidence and risk factors for exertional rhabdomyolysis in thoroughbred race horses in the United Kingdom. American Journal of Veterinary Research 60, 15626.
  • McKenzie E.C., Valberg S.J., Godden S.M., Pagan J.D., Carlson G.P., MacLeay J.M. & DeLaCorte F.D. (2002) Plasma and urine electrolyte and mineral concentrations in Thoroughbred horses with recurrent exertional rhabdomyolysis after consumption of diets varying in cation-anion balance. American Journal of Veterinary Research 63, 105360.
  • McKenzie E.C., Valberg S.J. & Pagan J.D. (2003) Nutritional management of exertional rhabdomyolysis. In: Current Therapy in Equine Medicine 5 (Ed. by N.E.Robinson), pp. 72734. Saunders, Philadelphia, PA.
  • McKenzie E.C., Valberg S.J., Godden S.M., Finno C.J. & Murphy M.J. (2004) Effect of oral administration of dantrolene sodium on serum creatine kinase activity after exercise in horses with recurrent exertional rhabdomyolysis. American Journal of Veterinary Research 64, 749.
  • McLean J.G. (1973) Equine paralytic myoglobinuria (azoturia): a review. Australian Veterinary Journal 49, 413.
  • Neuman R.J., Yuan B., Gerhard D.S., Liu K.Y., Yue P., Duan S., Averna M. & Schonfeld G. (2002) Replication of linkage of familial hypobetalipoproteinemia to chromosome 3p in six kindreds. Journal of Lipid Research 43, 40715.
  • Oki H., Miyake T., Hasegawa T. & Sasaki Y. (2005) Estimation of heritability for tying-up syndrome in the thoroughbred racehorse by gibbs sampling. Journal of Animal Breeding and Genetics 122, 28993.
  • Penedo M.C., Millon L.V., Bernoco D. et al. (2005) International Equine Gene Mapping Workshop Report: a comprehensive linkage map constructed with data from new markers and by merging four mapping resources. Cytogenetic and Genome Research 111, 515.
  • Raftery A.E. (1996) Approximate Bayes factors and accounting for model uncertainty in generalised linear models. Biometrika 83, 25166.
  • Raudsepp T., Gustafson-Seabury A., Durkin K. et al. (2008) A 4,103 marker integrated physical and comparative map of the horse genome. Cytogenetic and Genome Research 122, 2836.
  • Rossdale P.D., Hopes R., Digby N.J. & Offord K. (1985) Epidemiological study of wastage among racehorses 1982 and 1983. Veterinary Record 116, 669.
  • Shmulewitz D., Heath S.C., Blundell M.L. et al. (2006) Linkage analysis of quantitative traits for obesity, diabetes, hypertension, and dyslipidemia on the island of Kosrae, Federated States of Micronesia. Proceedings of the National Academy of Sciences of the United States of America 103, 35029.
  • Stanley R.L., McCue M.E., Valberg S.J., Mickelson J.R., Mayhew I.G., McGowan C., Hahn C.N., Patterson-Kane J.C. & Piercy R.J. (2009) A glycogen synthase 1 mutation associated with equine polysaccharide storage myopathy and exertional rhabdomyolysis occurs in a variety of UK breeds. Equine Veterinary Journal 41, 597601.
  • Swinburne J.E., Boursnell M., Hill G. et al. (2006) Single linkage group per chromosome genetic linkage map for the horse, based on two-three-generation, full-sibling, crossbred horse reference families. Genomics 87, 129.
  • Tozaki T., Mashima S., Hirota K., Miura N., Choi-Miura N. & Tomita M. (2001) Characterization of equine microsatellites and microsatellite-linked repetitive elements (eMLREs) by efficient cloning and genotyping methods. DNA Research 8, 3345.
  • Tozaki T., Hirota K., Hasegawa T., Tomita M. & Kurosawa M. (2005) Prospects for whole genome linkage disequilibrium mapping in thoroughbreds. Gene 346, 12732.
  • Tozaki T., Swinburne J., Hirota K., Hasegawa T., Ishida N. & Tobe T. (2007a) Improved resolution of the comparative horse-human map: investigating markers with in silico and linkage mapping approaches. Gene 392, 1816.
  • Tozaki T., Hirota K., Hasegawa T., Ishida N. & Tobe T. (2007b) Whole-genome linkage disequilibrium screening for complex traits in horses. Molecular Genetics and Genomics 277, 66372.
  • Wade C.M., Giulotto E., Sigurdsson S. et al. (2009) Genome sequence, comparative analysis, and population genetics of the domestic horse. Science 326, 8657.