Equine breeding selection has been developed by applying quantitative genetic methods for calculating the heritability of the complex traits such as performance in racing or sport competitions. With the great development of biotechnologies, equine molecular genetics has come of age. The recent sequencing of the equine genome by an international consortium was a major advance that will impact equine genomics in the near future. With the rapid progress in equine genetics, new applications in early performance evaluation and the detection of disease markers become available. Many new biomolecular tools will change management of horse selection, disease diagnosis and treatment. The purpose of this review is to present new developments in equine genetics and genomics for performance evaluation and health markers after a short summary of the previous knowledge about the genetic components of the exercise performance traits.
Equine breeding selection has been developed by applying quantitative genetic methods for calculating the heritability of the complex traits such as performance in racing or sport competitions. Genetic indices are now calculated routinely in different horse breeds in order to select the best stallions and mares according to a breeding plan with a well-defined genetic objective. Other exercise variables and disease characteristics have also been studied by the same quantitative genetic methods. With the great development of biotechnologies, the equine molecular genetics comes of age. The equine mitochondrial genome was sequenced in 1994 (Xu and Arnason 1994) and then partial genome sequencing was undertaken (Caetano et al. 1999; Milenkovic et al. 2002; Chowdhary et al. 2003). The recent equine genome sequencing by an international consortium (Wade et al. 2009) was a major advance that will impact equine genomics in the near future. With the rapid progress in equine genetics, new applications in early performance evaluation and the detection of disease markers become available. Many new biomolecular tools will change management of horse selection, disease diagnostics and treatment. In man, the research project HERITAGE family study provides continuous new knowledge and databases about the genes and quantitative trait loci (QTLs) related to exercise activity, health and performance. The last update of the gene list validated by experimental studies revealed 214 autosomal genes and quantitative trait loci and 7 others on the X chromosome (Bray et al. 2009). Moreover, there are 18 mitochondrial genes that have been shown to influence fitness and performance phenotypes. This important collection of genetic data related to exercise in man suggests that there will be great interest in studying equine genetics for performance evaluation and sports medicine applications.
The purpose of this review was to present new developments in equine genetics and genomics for performance evaluation and health markers after a short summary of previous knowledge about the genetic components of exercise performance traits.
Genetic component of the performance traits: how to measure the heritability of the exercise variables and performance traits?
With the development of equine exercise physiology, more functional traits can be measured for early evaluation of exercise ability. The objective is to predict the level of performance that a young horse can reach by measuring some physiological characteristics during an exercise test. For breeding purposes, the heritability of each trait should be determined to know if it is useful for breeding selection. This section presents how to calculate the heritability and apply the method to some physiological and locomotor traits related to the equine exercise ability. Several exercise traits have been studied in different breeds: muscle fibre types, heart rate, blood lactate, gait, jump style and conformation.
Heritability estimates (h2) of exercise traits have been calculated in order to estimate the influence of genetic components andenvironment effects (i.e. nongenetic) on trait variability. Heritability is the ratio between genetic variance and the total variance including environmental effects. For each variable, the variance components are calculated using restricted maximum likelihood resolution (REML). The genetic variance is estimated by the following variance analysis model including fixed (known during the experiment) and random effects (unknown: animal genetic effect and error). The total variability of the trait is explained in the sum of fixed effects (for example: age, height, track, breed) and the random animal genetic effect:
Heritability values (h2) >0.4 indicate that 40% of the trait could be transmitted to offspring by mare and stallion. Forty percent represents a high percentage of genetic influence. Performance heritability estimates in racing and equestrian sports are: 0.15–0.55 for flat gallop racing, 0.17–0.26 for trot or pace racing, 0.05–0.28 for showjumping and 3-day eventing and 0.11 for dressage (Hintz 1980; Langlois 1980; Ricard et al. 2000). In addition, it was interesting to note that there is no genetic correlation or a little negative correlation between jumping ability and other other sport performance and gait variables. Heritability estimates are valid only for the breed for which they were calculated and they should be recalculated periodically if there is an ongoing breeding programme in the population. The heritabilities of the important traits are used to calculate genetic index using best linear unbiaised predictor (BLUP) method in sport horses in different countries (Ricard et al. 2000).
Muscle power: percentages of fast and slow myosin heavy chains of gluteus medius
Myosin heavy chain (MHC) percentages were measured in small muscle samples of 20–30 mg collected by microbiopsy under local anaesthesia in Anglo-Arabian horses. The analysis of slow MHC I and fast MHC IIA and IIX was performed using monoclonal antibodies with an ELISA method (Barrey et al. 1999a). The heritability of the percentage of fast MHC was 0.13 in Anglo-Arabian and 0.28 in Andalusian horses (Rivero and Barrey 2001). The relative area h2 of the fast muscle fibres was 0.12 in Anglo-Arabian and 0.15–0.23 in Andalusian horses for fast and slow fibres, respectively. The capillary density h2 in Andalusian horses was 0.20. The relationships with performance should be taken into account: muscular traits are significantly different between high and poor performers in racing and jumping horses.
Evaluation of cardiac and aerobic capacity
French trotters performed an exercise test at increasing speed. Heart rate and blood lactate were measured in order to determine speed related changes of these variables. The speed eliciting a heart rate of 200 beats/min (V200) and a blood lactate concentration of 4 mmol/l (VLa4) were calculated. The V200 h2 was higher (0.46) than the VLa4 h2 (0.10), which was more influenced by the training effect (Barrey et al. 1999b). For harness trotters, VLa4 is less heritable but more related to performance than V200 (Leleu et al. 2003).
To measure kinetic and temporal variables of the walk, trot, gallop and jump the horses were equipped with the Equimetrix gait analysis system (Barrey and Galloux 1997; Barrey et al. 2001, 2002a). Two accelerometers were fixed with an elastic girth on the sternum ensuring that the transducer was held close to the horse's centre of gravity. The device recorded the dorso-ventral and longitudinal acceleration of the horse during the exercise and provided gait variables described in Tables 1 and 2. For the dressage gaits, the genetic effect of the breed was demonstrated in young horses of 3 European breeds: Hannover, Andalusian (PRE) and Selle Français (Barrey et al. 2002c). Heritability estimates of canter, gallop and jumping variables, such as take-off impulse were rather high: 0.23–0.52 (Barrey et al. 2002b; Barrey 2004). For equestrian sport horses, the use of gait variables, such as stride frequency, propulsion power and jump variables, is better indicated for early selection.
Table 1. Heritabilities (h2) of gait variables measured in Selle Français horses. Walk and trot were recorded in hand and canter under free condition. Canter heritabilities were higher than trot and walk heritabilities
Walk h2 (s.e.)
Trot h2 (s.e.)
Canter h2 (s.e.)
NA, Nonavailable data.
Mean propulsion vector
Table 2. Heritabilities (h2) of jump variables measured in Selle Français horses and Poney de Selle. The jumps were recorded under free condition. Jump heritabilities were higher than walk and trot heritabilities
Take off force
Using a 2D image analysis, it has been possible to measure French saddle horses and to calculate the heritability estimates of the main morphological traits. Conformation is poorly heritable in Selle Français horses and more heritable in Poney de Selle (Barrey et al. 2002a). In the 2 breeds, these conformation traits are poorly related to the sport performances. However, in a review of the conformation h2 obtained by other methods (judge scoring and direct measurements) and in other breeds, the heritabilities could be higher for some (Saastamoinen and Barrey 2000).
New applications in equine genomics available after the recent genome publication
Briefly, the equine genome includes 20,322 predicted protein-coding genes according to the annotation process made by alignments and comparisons with the human, mouse and dog genomes. Some transcriptomic analysis showed that 87% of 18,039 genes can be detected using expression microarrays designed with equine probes. In order to have markers throughout the equine genome, a single nucleotide polymorphism (SNP) map was designed including a of total 106 SNP candidates with an average density of 1 SNP/2 kb. It was possible to discriminate 11 different breeds by genotyping a set of 1007 SNPs. This important equine genetic result allow production of new genetic investigations in various domains and design new genomic tools such as SNP chips (Equine SNP50)1 and expression microarrays2 to analyse gene expression and regulation.
Looking for genetic markers of performance, health and disease using a genome-wide scan
Using a genome-wide scan study with microsatellites markers or more recently SNP chips, it is now possible to identify some interesting genes linked to some performance traits or health troubles such as osteochondrosis (OC). Some small variations of the gene sequences such as a SNP can be detected and the association to interesting traits can be demonstrated by statistical tests. Then, if the genetic marker is close to an interesting candidate gene, it could be a way to discover the genetic determinism of a particular phenotype. Another application is to use the best informative SNP markers to genotype a subject in relation to a specific phenotype.
Because selection has been recent and intense in a closed population that stems from a small number of founder animals, Thoroughbreds represent a unique population, within which to identify genomic contributions to exercise-related traits. In a wide genome scan study of 112 Thoroughbred horses, Gu et al. (2009) identified some candidate racing performance genes by using 394 microsatellite markers. These genes were close to the genetic markers and were identified to be involved in energetic metabolism (alcohol dehydrogenase, ADHFE1; pyruvate dehydrogenase kinase 4, PDK4), mitochondrial respiration (mitochondrial fission regulator 1, MTFR1; COX cytochrome units), hypoxia (HIF1A), fatty acid oxidation (acyl-CoA synthetase short-chain family member 1, ACSS1), phosphoinositide-mediated signalling, insulin receptor signalling (PI3KR1), and muscle strength and integrity actin alpha 1, actinin alpha 2 and tenascin C genes (ACTA1, ACTN2, TNC).
Osteochondrosis is a major locomotor disorder in sport and racing horses. Several projects were started to detect genetic markers of this multifactorial joint disease affecting young horses. In a first study, Dierks et al. (2007) performed a genome-wide scan to detect QTLs for OC and osteochondrosis dissecans (OCD) in horses. The genetic marker set comprised 260 microsatellites. Data were collected from 211 Hanoverian Warmblood horses consisting of 14 paternal half-sib families. Traits used were OC (fetlock and/or hock joints affected), OCD (fetlock and/or hock joints affected), fetlock OC, fetlock OCD, hock OC and hock OCD. Genome-wide significant QTLs were located on equine chromosomes 2, 4, 5 and 16. QTLs for fetlock OC and hock OC partly overlapped on the same chromosomes, indicating that these traits may be genetically related. QTLs reached the chromosome-wide significance level on 8 different equine chromosomes: 2, 3, 4, 5, 15, 16, 19 and 21. This whole-genome scan was a first step toward the identification of candidate genome regions harbouring genes responsible for equine OC. In a second study, Lampe et al. (2009a) used 29 microsatellite markers distributed in chromosome 5 to refine the map positions of the QTL already identified for OC in the fetlock. Collagen type XXIV alpha 1 was identified as a potential functional candidate gene for the fetlock OC. In another chromosome-wide linkage analysis, Lampe et al. (2009b) revealed another QTL on chromosome 18 for OC in fetlock, hock or both joints, as well as for OCD in hock joints. Within this QTL for equine OC, the parathyroid hormone 2 receptor gene could be identified as a positional candidate gene. These last results are a further step toward the identification of genes responsible for OC in horses.
Using a new SNP chip (Equine SNP50)1, which includes 54,602 SNP markers distributed among the whole equine genome, other genome-wide scan projects are ongoing to detect SNP markers of showjumping and endurance race ability. Other projects are dealing with OC and larynx paralysis in French saddle horses and French Trotters.
Detecting the genes of performance, health and disease using genotyping
Using the genome database, comparative genomics and genome-wide scan results, it is now possible to identify some interesting genes variability linked to some performance traits such as gallop distance ability or health troubles such as polysaccharide storage myopathy (PSSM). Some small variations of the gene sequences such as a SNP can be related to the change of function of the gene in case of a mutation. By genotyping one identified SNP it is possible to make a genetic diagnosis of a disease such as PSSM or to predict a part of the performance trait such as the distance ability in racing.
Polysaccharide storage myopathy is an equine glycogenesis (metabolic myopathy) related to an abnormal glycogen accumulation. In man, 11 types of glycogenesis have been described, each containing several subtypes due to various gene mutations (Yoon 2006). In horses, it was first described in Quarter Horses (Valberg et al. 1992) and then in other breeds: Spanish Purebred horses, Belgian Draught horses, Morgan, Arabian, Standardbred, Warmbloods, ponies and mules (Quiroz-Rothe et al. 2002; Valentine and Cooper 2005; McCue et al. 2006, 2009; McCue and Valberg 2007). The common clinical signs of PSSM are poor performance, back soreness, exercise intolerance, gait disorders, generalised muscle atrophy, rhabdomyolysis, and spontaneous decumbency with inability to rise and episodic ‘colic’. A PSSM phenotype can be characterised by histological demonstration in striated muscle fibres of an accumulation of some periodic acid Schiff-positive amylase-resistant polysaccharides appearing ultrastructurally as glycogen-like particles (Firshman et al. 2006). In horses, a specific form of fatal glycogen storage disease was observed in a Quarter Horse foal where a nonsense mutation in codon 34 of the glycogen branching enzyme (GBE1) was identified (Ward et al. 2004). This PSSM form can be compared to human glycogenosis type IV where several mutations and deletions explain the total or partial deficiency of the GBE1 gene (Yoon 2006). In mature horses, a G-to-A mutation within the gene encoding glycogen synthase 1 (GYS1) has recently been observed in Quarter Horses and some Warmbloods (McCue et al. 2008; Herszberg et al. 2009). The PSSM mutation was identified in horses from 17 different breeds. The prevalence of the GYS1 mutation in PSSM horses was high in draught- (87%) and Quarter Horse-related breeds (72%) and lower in Warmbloods (18%) and other light horse breeds (24%), when diagnosis was based on grade 2 diagnostic criteria. Overall, the PSSM mutation was present in 16% of grade 1 and 70% of grade 2 PSSM horses. False-positive diagnosis, as well as the possibility of a second glycogenesis in horses with neuromuscular disease, might explain the absence of the GYS1 mutation in horses diagnosed with excessive glycogen accumulation in muscle. In some severe cases of PSSM, another gene mutation has been observed in the ryanodine receptor RYR1 that is responsible for an excess calcium release by the sarcoplasmic reticulum into the cytosol of the contracting muscle fibre (McCue et al. 2009). In addition to the clinical examination and muscular histology diagnosis, the genotyping of the genes GYS1 and RYR1 could be used to confirm the PSSM disease to establish a risk evaluation and to avoid the gene transmission to offspring.
Another example of genotyping in racing performance evaluation involves variants of the myostatin gene (MSTN), which are associated with muscle hypertrophy phenotypes in a range of mammalian species, most notably cattle, dogs, mice and man. Using a sample of registered Thoroughbred horses, Hill et al. (2010) have identified a novel MSTN sequence polymorphism that is strongly associated with best race distance among elite racehorses. The authors observed that the SNP C/C horses are suited to fast, short-distance races, C/T horses compete favourably in middle-distance races, and T/T horses have greater stamina for longer distance (Figure 2). Evaluation of retrospective racecourse performance (n = 142) and stallion progeny performance predict that C/C and C/T horses are more likely to be successful 2-year-old racehorses than T/T animals. This is the first study to demonstrate a relationship between a racing performance trait and a SNP gene variability.
Investigation of the phenotype by the analysis of gene expression and regulation
Gene expression and regulation determine most of the phenotype. The analysis of the cell transcriptome using DNA microarray provides qualitative and quantitative information about the up- and downregulated pathways involved during exercise. The collected data are massive and complex and should be studied by sophisticated data mining methods in order to extract only the informative data. The first transcriptomic studies in exercising horses were performed in the blood leucocytes of performers and disqualified endurance horses (Barrey et al. 2006) and in muscles of Thoroughbred racehorses in training (McGivney et al. 2009).
Barrey et al. (2006) analysed the gene expressions by using DNA microarrays in order to study the physiological adaptations and metabolic disorders in endurance horses. In order to show that genes are modulated in leucocytes in relationship with performance and clinical status of the horses, gene expression in leucocytes, haematological and biochemical parameters were compared between successful and disqualified endurance horses. Blood samples were collected at rest and just after a 140–160 km endurance race in 2 groups of horses: 10 continuing successfully and 10 disqualified at a vet-gate for metabolic disorders. Total RNA was extracted from the total blood cells, checked for purity, amplified and hybridised using mouse cDNA microarrays including 15,264 unique genes. Differential gene expressions were studied by hybridisation of each sample collected at the end of the race vs. the control sample collected before the race. Some significant differences were observed in the haematology and biochemistry of the 2 groups. In the disqualified group, rhabdomyolysis was confirmed with CK 13,124 u/l and AST 1242 u/l. The list of 726 (including 603 annotated genes) significant genes was filtered according to a high P value cut-off (P<0.00001). Among them, 130 were upregulated (expression ratio >1.5) and 288 were downregulated (<1/1.5). Analysis of variance revealed 62 genes differentially expressed (P<0.05) in disqualified and successful horses. The expression levels of 28 and 50 genes were significantly correlated (r2>0.75) with CK and AST levels in the disqualified group, respectively. The gene ontology classification showed that more genes were upregulated in the successful than in the disqualified horses. More genes were downregulated in the disqualified horses. Long exercise induced many significant gene modulations in leucocytes and many of them are involved in inflammation signalling. Some genes were expressed in relationship with the clinical phenotype observed in the disqualified horses: inflammation, rhabdomyolysis and haemolysis. Some of these genes could be interesting markers and/or candidates of poor performance or pathologies.
McGivney et al. (2009) suggested that exercise stimulates immediate early molecular responses as well as delayed responses during recovery, resulting in a return to homeostasis and enabling long-term adaptation. Global gene expression changes in equine skeletal muscle following exercise have not been reported. Therefore, to identify novel genes and key regulatory pathways responsible for exercise adaptation the authors used equine-specific cDNA microarrays to examine global mRNA expression in skeletal muscle from a cohort of Thoroughbred horses (n = 8) at 3 time points following a single bout of treadmill exercise. Skeletal muscle biopsies were taken from the gluteus medius before exercise, immediately after and 4 h after exercise. While only 2 genes had increased expression after exercise (P<0.05), by 4 h after exercise, 932 genes had increased (P<0.05) and 562 genes had decreased expression (P<0.05). Functional analysis of genes differentially expressed during the recovery phase (4 h after exercise) revealed an over-representation of genes localised to the actin cytoskeleton and with functions in the MAPK signalling, focal adhesion, insulin signalling, mTOR signalling, p53 signalling and type II diabetes mellitus pathways. After the exercise, an over-representation of genes involved in the stress response, metabolism and intracellular signalling was observed. These findings suggest that protein synthesis, mechanosensation and muscle remodelling contribute to skeletal muscle adaptation towards improved integrity and hypertrophy. This was the first study to characterise global mRNA expression profiles in equine skeletal muscle using an equine-specific microarray platform. Novel genes and mechanisms are revealed that are temporally expressed following exercise in the equine skeletal muscle transcriptome.
More recently, muscle diseases such as PSSM (Barrey et al. 2009) and recurrent exertional rhabdomyolysis (RER) were investigated using transcriptomic analysis (Jayr 2009). Several cases of myopathies have been observed in the Cob Normand breed, with muscle histology examinations revealing that some families suffer from PSSM. It is assumed that a gene expression signature related to PSSM should be observed at the transcriptional level because the glycogen storage disease could also be linked to other dysfunctions in gene regulation. Therefore, the functional genomic approach could be conducted in order to provide new knowledge about the metabolic disorders related to PSSM. It was proposed to explore the PSSM muscle fibre metabolic disorders by measuring gene expression in relationship with the histological phenotype. Genotypying analysis of GYS1 mutation revealed 2 homozygous (AA) and 5 heterozygous (GA) PSSM horses. In the PSSM muscles, histological data revealed positive amylase resistant abnormal polysaccharides, inflammation, necrosis, lipomatosis and active regeneration of fibres. Ultrastructural evaluation revealed a decrease of mitochondrial number and structural disorders. Extensive accumulation of an abnormal polysaccharide displaced and partially replaced mitochondria and myofibrils. The severity of the disease was higher in the 2 homozygous PSSM horses. Gene expression analysis revealed that 129 genes were significantly modulated (P<0.05). The following genes were upregulated >2 fold: IL18, CTSS, LUM, CD44, FN1, GST01. The most down-regulated genes were the following: mitochondrial tRNA, SLC2A2, PRKC alpha, VEGF alpha. Data mining analysis showed that protein synthesis, apoptosis, cellular movement, growth and proliferation were the main cellular functions significantly associated with the modulated genes (P<0.05). Several upregulated genes, especially IL18, revealed a severe muscular inflammation in PSSM muscles. The upregulation of glycogen synthase kinase-3 (GSK3B) in its active form could be responsible for glycogen synthase (GYS1) inhibition and hypoxia-inducible factor (HIF1A) destabilisation (Figure 3). The main disorders observed in PSSM muscles could be related to mitochondrial dysfunctions, glycogenesis inhibition and the chronic hypoxia of the PSSM muscles.
Recently, microRNAs (miRNAs) were detected in muscle and blood and could be interesting markers of diseases in the near future, as has already been presented in human cancer prognostics (Calin et al 2006). MicroRNAs are small, endogenous, noncoding, interfering RNA molecules of 18–25 nucleotides that are regarded as major regulators in eukaryotic gene expression. They play a role in developmental timing, cellular differentiation, signalling and apoptosis pathways. Because of the central function of miRNAs in the proliferation and differentiation of the myoblasts demonstrated in mice and man, it is assumed that they could be present in equine muscles and their expression profile may be related to the muscle status. Several miRNA candidates (miRNA-1, 23, 30, 133, 181, 188, 195, 206, 339, 375) were significantly detected in the normal and pathological muscles and some other miRNAs in blood samples collected in PSSM cases (Barrey et al. 2010). This first study on muscular miRNAs profiles in equine myopathies indicated that it is possible to discriminate pathological from control horses according to their miRNA profile. The RER miRNA profile was more specific and contrasted than the PSSM miRNA profile.
Conclusion: genetic and genomics perspective in equine exercise
With the great development of the genomic technologies many new applications will soon be available for the horse industry: candidate genes responsible for diseases and performance, genotyping, diagnostic tools and treatments. The first genotype test for racing distance ability is now available for flat racing early evaluation. The diagnostic of PSSM myopathy by genotyping is routinely applied. Within 5 years, other genotyping methods will be available for early performance evaluation and genetic diseases diagnostic and risk evaluation like in osteochondrosis. Perhaps new genomic treatments will be designed to treat some myopathies and other metabolic diseases as are now applied in man. The drawback of the biotechnologies’ availability will be illicit use for doping using muscular transgenesis for example. Hopefully, research projects are already going on to detect this new type of genomic doping both in man and horses.
Mrs Veronica Blin (INSERM, UBIAE, U902) is greatly acknowledged for her English revision.
Conflicts of interest
The author declares no potential conflicts.
1 Illumina, San Diego, California, USA.
2 Agilent technologies, Santa Clara, California, USA.