Article Text

This article has a correction. Please see:

Download PDFPDF

Genomics of elite sporting performance: what little we know and necessary advances
  1. Yannis Pitsiladis1,
  2. Guan Wang1,
  3. Bernd Wolfarth2,
  4. Robert Scott3,
  5. Noriyuki Fuku4,
  6. Eri Mikami4,
  7. Zihong He5,
  8. Carmen Fiuza-Luces6,
  9. Nir Eynon7,
  10. Alejandro Lucia6
  1. 1College of Medicine, Veterinary and Life Sciences, Institute of Cardiovascular and Medical Sciences, University of Glasgow, Glasgow, Lanarkshire, UK
  2. 2Department of Preventive and Rehabilitative Sports Medicine, Technical University Munich, Munich, Germany
  3. 3MRC Epidemiology Unit, Institute of Metabolic Science, Cambridge, UK
  4. 4Department of Genomics for Longevity and Health, Tokyo Metropolitan Institute of Gerontology, Tokyo, Japan
  5. 5Biology Center, China Institute of Sport Science, Beijing, China
  6. 6School of Doctorate Studies and Research, European University of Madrid, Madrid, Spain
  7. 7Institute of Sport, Exercise and Active Living (ISEAL), Victoria University, Melbourne, Australia
  1. Correspondence to Dr Yannis Pitsiladis, College of Medicine, Veterinary and Life Sciences, Institute of Cardiovascular and Medical Sciences, University of Glasgow, Glasgow, G12 8QQ, Scotland; Yannis.Pitsiladis{at}Glasgow.ac.uk

Abstract

Numerous reports of genetic associations with performance-related phenotypes have been published over the past three decades but there has been limited progress in discovering and characterising the genetic contribution to elite/world-class performance, mainly owing to few coordinated research efforts involving major funding initiatives/consortia and the use primarily of the candidate gene analysis approach. It is timely that exercise genomics research has moved into a new era utilising well-phenotyped, large cohorts and genome-wide technologies—approaches that have begun to elucidate the genetic basis of other complex traits/diseases. This review summarises the most recent and significant findings from sports genetics and explores future trends and possibilities.

  • Genetics/sex testing
View Full Text

Statistics from Altmetric.com

Introduction

Despite numerous attempts in recent years to discover genetic variants associated with elite athletic performance and, more specifically, elite/world-class athletic status, there has been limited progress owing to few coordinated research efforts involving major funding initiatives/consortia and the reliance on candidate gene analyses, involving a small number of single nucleotide polymorphisms (SNPs) and structural variants (eg, the commonly studied insertion/deletion polymorphisms). Nevertheless, over 200 SNPs associated with physical-performance traits, and over 20 SNPs associated with elite athletic status, have been reported in the literature and summarised on a yearly basis in the ‘The Human Gene Map for Performance and Health-related Fitness Phenotypes’ until 2009.1 Owing to the massive increase in related papers, the authors have changed the format summarising only the key findings of each year in ‘Advances in Exercise, Fitness, and Performance Genomics’2–5 series. However, most reported associations are cited in studies with small sample sizes without robust replication, and therefore are most likely type 1 errors. It is widely acknowledged that there will be many genes involved in physical performance phenotypes, and hence it is timely that genetic research has moved to the genomics era, that is, the simultaneous testing of multiple genes is now possible. New approaches involving large, well-funded consortia and utilising well-phenotyped large cohorts and genome-wide technologies will be necessary for meaningful progress to be made. This review summarises the most recent and significant findings from sports genetics and explores future trends and possibilities.

Study designs, strategies and methodologies

Twin and family studies

Similar to other areas of research, family or twin studies were initially the main focus investigating the genetic basis of human performance. The initial studies in the 1970s assessed indirectly the genetic basis of human performance using twin models and comparing the intra-pair variation between monozygotic (MZ) and dizygotic (DZ) twins, a concept referred to as the heritability estimate (h2), which reflects the population variance in a trait attributable to genetic factors (assuming a simple additive model of heredity) and is calculated by dividing the difference of the variance between DZ and MZ twins by the variance of DZ twins. When this approach was applied to maximal oxygen consumption (Embedded Image max) in 25 pairs of twins (15 MZ and 10 DZ preadolescent boys), high heritability estimates were reported (eg, h2=93.4%).6 In other words, genetics could explain as much as 93.4% of the phenotypic variation in Embedded Image max. Similarly, high heritability estimates were reported in 15 MZ and 16 DZ twins of both genders (h2=96.5%)7 for the variation in skeletal muscle fibre composition. Genetic influences on other performance-related attributes such as body composition and motor activities (eg, walking, running, throwing, balancing) as well as training-induced improvements in Embedded Image max were also reported using similar methods.8–12 A more contemporary view based on results from the HERITAGE Family Study is that genetic factors can explain ∼50% of Embedded Image max when adjusted for age, body mass and body composition as covariates.13 Notably, in a study of 4488 adult British female twins, the heritability of athlete status was estimated to be at 66%.14 Given these exceptionally high heritability estimates, this concept has received considerable criticism with arguments that the high heritability estimates are related to low twin numbers and the near identical social environment of the studied twins.15 ,16 Recent studies also report high heritability estimates for neuromuscular performance and body composition.17–19

Despite the suggestion of significant heritable components for a range of performance-related traits, family-based studies do not offer insight into the specific genetic variation underlying these heritable components. The limitations and criticisms of the early indirect methods required the focus to be shifted to the continuously developing molecular-based laboratory methods to test directly the interaction between genetic and environmental factors, not only in family or twin studies but also in populations of interest for complex traits; for example, elite athletic status, where interindividual phenotypic variations are caused by a heterogeneous, polygenic model with multiple gene variants involved and interacting with environmental factors.1 ,20 ,21

Genome-wide association studies: hypothesis-free approach

A number of different methodological approaches within the field of genetic epidemiology have been utilised to unravel the genetic basis of elite human performance. Owing to the development of more advanced gene discovery techniques, genetic studies are no longer restricted to family/twin studies but expanded to include the assessment of genetic variants (ie, mostly SNPs) within a population of interest. Population-based case-control studies are being used extensively at present, and can be further differentiated into hypothesis-free (sometimes referred to as a ‘fishing trip’, ie, no assumptions made about the genomic location of associated variants) and the more commonly used hypothesis-driven approaches, where the search for association may be restricted to particular genes of interest.

Advances in molecular technologies have enabled researchers to apply genome-wide association (GWAS) approaches to the field. GWAS examines the ‘association of genetic variation with outcomes or phenotypes of interest by analysing 100 000 to several millions of SNPs across the entire genome without any previous hypotheses about potential mechanisms’. GWAS has been successful in identifying novel genetic variants for age-related macular degeneration,22 type 2 diabetes mellitus,23 the interleukin 23 pathway in Crohn’s disease24 and obesity-related traits.25 This promising approach is not without important limitations. For example, human height is a highly heritable quantitative trait (up to 90% of population variance)26–30 as well as stable and easy to measure, although one of the largest studies to date (n=183 727) identified at least 180 loci associated with adult height, but together these explained only 10% of the variation in height. This is in large part related to the small effect size of most of these genetic variants. Furthermore, genetic variants associated with most complex diseases do not show predictive utility.31 Thus, much of the heritability of complex traits is missing,32 and numerous explanations have been proposed to account for this missing heritability. There have been suggestions that common variants do explain up to 45% of the variance in height,33 but the small effect size of these variants may render these variants undetectable by common study designs.32 Despite several limitations, these studies confirm that GWAS is able to detect many loci that implicate biologically related genes and pathways.34 The occurrence of rare variants which are not captured by GWAS may partly explain this limited success in determining the genomics of adult height. Following GWAS, additional approaches, such as fine mapping and sequencing, may be used to find common SNPs with larger effect sizes than GWAS tagging SNPs or identify rarer variants across GWAS loci. The hypothesis-free GWAS design is the most popular of the current widespread approaches as it allows one to (1) detect smaller gene effects by narrowing down the genomic target region precisely with new chips; (2) maximise the amount of variation captured per SNP with a fixed set of markers and (3) reduce genotyping costs, which make this approach attractive.35 Other factors relating to the sample population (family history of traits, ethnically homogeneous populations and a sample size of at least several thousands) and differences in statistical approaches (eg, the conservative Bonferroni correction for multiple testing vs the non-conservative false discovery rate correction or the application of permutation testing approaches) need to be carefully considered to ensure successful application of the GWAS approach.36 GWAS of elite human athletic performance are ongoing,37–40 but there are no published papers to date.

Candidate gene analysis: hypothesis-driven approach

The most extensively used candidate gene association study approach requires a prior hypothesis that particular genes of interest contain variants that may be associated with a trait or disease. Typically, variants in a gene or genes of interest are genotyped in cases and controls, or assessed for association(s) with quantitative trait(s). This approach is effective in detecting genetic variants with a small or modest influence on common disease or complex traits. Functional SNPs with tag SNPs (by use of linkage disequilibrium) which would cover the entire candidate gene have been used in many candidate gene association studies.35 However, in this approach, candidate genes ought to be selected if there is good evidence that (1) the proposed candidate gene is biologically relevant to the main phenotype/complex trait of interest (eg, physical performance/aerobic capacity, adiposity); (2) the variants of the candidate gene influence the overall function of the gene (eg, variation in physiological ACE inhibitors activity levels are linked to polymorphisms in the ACE gene—see section on ‘Genes and polymorphisms with reasonable replication’) and (3) the polymorphisms of the selected candidate gene are frequent enough in the population to allow meaningful statistical analysis (eg, typical allele frequencies for the I and D alleles of the ACE I/D polymorphism in a European population are ∼43% and 57%, respectively). When these criteria are not fulfilled and candidate genes are selected based primarily on the interest of the research group, this approach generates conflicting results with low statistical power and difficulty of being replicated in other populations, and thus low validity.41 In addition, the candidate gene approach has largely been implemented in studies with small participant numbers and often without robust replication, perhaps partly attributable to publication bias.

Major study cohorts

The genotyping of athletes of the highest performance calibre such as world record holders, world champions and Olympians is desirable and may circumvent the need for very large athlete cohorts in order to discover performance-associated polymorphisms. The number of large genetic cohorts of world-class athletes from a variety of countries and sports with extensive physical performance phenotypes is limited. The following are the most significant elite athlete cohorts based on current publication outcomes.

Genathlete study

In the Genathlete study, a classical case–control study, the endurance of athletes with a high Embedded Image max was compared with that of control participants with a low to average Embedded Image max.42–44 Data were analysed using the candidate gene approach, which allowed the distribution of particular genetic variants with respect to the phenotype Embedded Image max in both groups to be assessed. As this type of assessment requires large participant numbers, this study was designed as a multicentre study. To exclude influences owing to a regionally different distribution of genetic variants, particular attention was paid to a comparable distribution of the regional origin of participants. Currently, this cohort involves more than 600 participants (∼300 athletes and ∼300 controls) and therefore constitutes one of the largest matched case–control studies in this field.43 ,44

Elite Russian athlete cohort

One of the largest studies of elite athletes involves elite Russian athletes from mixed athletic disciplines.45–47 In the most recent study, 998 male and 425 female Russian athletes of regional or national competitive standard were recruited from 24 different sports.47 Athletes were stratified into five groups according to event duration (very long-endurance, long-endurance and middle-endurance), mixed ‘anaerobic/aerobic’ activity group and power group (predominantly anaerobic energy production).

Elite East African athlete cohorts

The phenomenal success of athletes from Ethiopia and Kenya in endurance running events is well recognised. Middle-distance and long-distance runners from Ethiopia and Kenya hold over 90% of both the all-time world-records and the current top-10 positions in world event rankings.48 Moreover, these successful athletes come from localised ethnic subgroups within their respective countries.49 ,50 In order to investigate the East African running phenomenon, a first study50–52 involved 76 endurance runners from the Ethiopian junior-level and senior-level national athletics teams (12 women and 64 men), 315 controls from the general Ethiopian population (34 women and 281 men), 93 controls from the Arsi region of Ethiopia (13 women and 80 men), and 38 sprint and power event athletes from the Ethiopian national athletics team (20 women and 18 men). A similar approach was conducted in a study53 with 291 elite Kenyan endurance athletes (232 men) and 85 control participants (40 men). Seventy of the athletes (59 men) had competed internationally representing Kenya and achieved remarkable success.

Elite Jamaican and USA sprint cohorts

These cohorts are comprised of elite Jamaican and African-American athletes representing the highest level of sprinting performance and geographically matched controls. In the Jamaican cohort, 116 athletes (men=60 and women=56) and 311 control participants (throughout the whole island; men=156 and women=155) were recruited.54 A total of 71 and 35 athletes had participated in 100–200 m and 400 m sprint events, respectively; and 10 athletes were involved in the jump and throw events. These athletes can be further classified into national (n=28) and international athletes (n=88) who were competitive at the national level in Jamaica and the Caribbean or at major international competitions for Jamaica. Among the 88 international athletes, 46 had won medals at major international events or held world records in sprinting. In the African-American cohort, samples from 114 elite sprint athletes (men=62 and women=52) and 191 controls (throughout the USA; men=72 and women=119) were collected.54 Among these athletes, 48, 42 and 24 athletes participated in 100–200 m, 400 m and jump and throw events, respectively. Athletes can be subdivided into 28 national and 86 international athletes; 35 of these athletes had won medals at international games and/or broken sprint world records.

Elite Australian athlete cohort

Australia has provided valuable genetic information on elite sprinters and endurance performers. The cohort comprises 429 elite athletes from 14 different sports and 436 unrelated controls. A subgroup of 107 and 194 participants were classified as elite sprinters and endurance runners, respectively.55 This cohort was studied to postulate, for the first time, the ACTN3 gene as a strong candidate to influence elite athletic performance (to be discussed in the ‘Genes and polymorphisms with reasonable replication’ section).

Elite Japanese athlete cohort

Japanese athletes are successful in international competitions such as the Olympics, especially in endurance-oriented events such as the Marathon and swimming events. This cohort is comprised of 717 elite Japanese athletes and 814 controls. Athletes are either national (participants in national competitions) or international athletes (participants in the Olympic Games, World and Asian Championships), including several medallists at these international games and world record holders. This Japanese athlete cohort comprises 381 track and field athletes, 166 swimmers and 170 Olympians from various sports. This cohort was initially established in order to identify both nuclear and mitochondrial DNA polymorphisms/haplogroups associated with elite Japanese athlete status and performance-related traits.56–61

Elite European and Asian swim cohort

Two elite swim cohorts, comprising Caucasian and East Asian swimmers, respectively, have been established. The Caucasian cohort comprised 200 elite Caucasian swimmers from the European, Commonwealth, American and Russian subcohorts. Swimmers were categorised as short and middle distance (≤400 m, n=130) or long distance swimmers (>400 m, n=70). Caucasian swimmers were all highly competitive and of world-class status having represented their countries in international competitions. Caucasian controls were drawn from a previous published report.62 Elite Japanese (n=158) and Taiwanese (n=168) swimmers were recruited and classified as short distance (<200 m, n=166) and middle distance (200–400 m, n=160), and none of these Asian swimmers competed at a distance greater than 400 m. East Asian swimmers were world-class having also competed in international competitions such as the Olympics, World Championships and Asian Games, or were competitive in national competitions. Controls were pooled from general Japanese (n=649) and Taiwanese (n=603) populations, and were healthy adults of both sexes and not professionally connected with athletics/sport. Two candidate genes (ACE inhibitors and α-actinin-3 (ACTN3); see next section) have been studied to date in this cohort.63

Spanish cohort

Extensive researches have been conducted on Spanish male athletes, with the most representative cohort comprising endurance world-class athletes (n=100, including 50 Olympic-class endurance runners and 50 professional cyclists (most of whom are Tour de France finishers, including stage winners))64 ,65 and world-class rowers (n=54, lightweight category, most of whom are medallists in world championships).66 This cohort also includes the majority of all-time best Spanish judo male athletes (n=108),67 elite swimmers (n=88)68 and track and field elite power athletes (n=53).69

Elite Israeli cohort

This cohort is comprised of 74 endurance and 81 sprint/power male athletes, who are current and former track and field athletes, as well as 240 matched controls. Athletes were carefully selected and included only if their main events were the 10 000 m run or the marathon (endurance group); and only if their main events were the 100–200 m dash and long-jump (sprint/power group). According to their personal best, athletes were further divided into two subgroups—the elite-level (those who had represented Israel in track and field world championships or in the Olympic Games) and the national-level.

Chinese cohort

He et al70 ,71 have recently conducted research in a Chinese cohort (of Han origin) made up of the best endurance runners (from 5000 m to marathon) of this country (current total n=241, 118 men). An important novelty of research on this Chinese cohort is the replication of results in a different (Caucasian) group of athletes and especially the analysis of the functionality of the SNPs, which was found to be associated with elite endurance status (using a dual-luciferase reporter assay).70

Genes and polymorphisms with reasonable replication

ACE inhibitors and the renin-angiotensin-aldosterone-system

One of the most widely studied candidate genes for athletic performance is the ACE inhibitors gene. ACE is a peptidase known to regulate blood pressure by catalysing the conversion of angiotensin I to the vasoconstrictor angiotensin II and also degrading the vasodilator bradykinin.72 Interindividual variation in physiological ACE activity levels has been linked to polymorphisms in the ACE gene. Notably, ACE insertion-deletion (I/D) (rs4340), in which ‘I’ refers to the presence (insertion) and ‘D’ to the absence (deletion) of a 287 bp sequence in an Alu sequence of intron 16 in the ACE gene at chromosome location 17q23, can account for up to 47% of ACE activity variance in participants (ie, Caucasians and Asians, not Africans) with an additive effect across the II, ID and DD genotypes.72 Regarding physical performance, ACE I/D genotypes have been associated with a wide range of phenotypes. ACE II participants show significantly higher muscle efficiency gains from training than DD individuals,73 as well as greater improvements in running economy, or ability to sustain submaximal pace with lower oxygen consumption.74 Additionally, the I-allele associates with muscular endurance gains from training,75 which may relate to higher type 1 (slow-twitch) muscle fibre preponderance in ACE II participants.76 In terms of overall sporting ability, the I-allele has been associated with superior performance in British mountaineers,75 South African triathletes,77 British distance runners78 and Australian rowers.79 In contrast, the D-allele has been associated with success in power-oriented sports such as short-distance swimming63 ,80 and sprinting.78 Notably, a recent study63 reported that the ACE D-allele was associated with short and middle distance swimmer status in Caucasian swimmers, whereas the ACE I-allele was found to be overrepresented in East Asian short distance swimmers. Although this finding might be explained by the different risk alleles being responsible for the associations in swimmers of different ethnicities, it requires to be further confirmed in future studies. Other inconsistencies in the literature regarding ACE findings also exist. For instance, the D-allele has been found to be both positively81 and negatively82 associated with Embedded Image max. A study of 230 elite Jamaican and American sprinters found no association of either allele with sprint athlete status.54 A cohort of 192 athletes of mixed Caucasian nationalities and endurance sporting disciplines did not exhibit I-allele frequencies, which were significantly different from geographically-matched controls, and nor did the I-allele frequency associate with Embedded Image max in these athletes.43 Several other studies involving Caucasian populations also found no association between the I-allele and elite physical performance,45 ,62 ,83 and an opposite finding (ie, D-allele associated with endurance performance) in Israeli84 and Korean85 cohorts. Despite the many inconsistencies in replication, the ACE gene remains as a candidate possibly influencing elite performance.

α-Actinin-3

α-Actinin-3 is an actin-binding protein and a key component of the sarcomeric Z-line in skeletal muscle. Expression of ACTN3 (at 11q13.1) is limited to type II (ie, fast, mostly glycolytic) muscle fibres which can generate more force at high velocity. Homozygosity for the common nonsense polymorphism R577X (rs1815739) in the ACTN3 gene results in deficiency of ACTN3 in a large proportion of the global population.86 Yang et al87 examined the R577X genotype frequencies in three African populations (Kenya, Nigeria and Ethiopia) in comparison with non-African populations (Europe, Asia and Australia). Extremely low 577XX genotype frequencies were observed in Kenyan and Nigerian athletes versus controls (1% vs 1% and 0% vs 0%, respectively), and they were much lower than in any other non-African populations, that is, frequency of the 577XX genotype of 18% in Australian Caucasians, 10% in Aboriginal Australians, 18% in Spanish Caucasians and 25% in Japanese). These results also implied that the ACTN3 deficiency was not a major influence on performance in African athletes. This polymorphism does not appear to result in pathology, although it could alter muscle function.88–92 Furthermore, a strong association has been reported between the ACTN3 R577X polymorphism and elite athletic performance in Caucasian populations.55 ,93–98

The 577XX genotype was found at a lower frequency in elite Australian sprint/power athletes relative to controls,55 and this finding was replicated in the Finnish,96 Greek,97 Israeli99 and Russian athletes.94 In particular, in a study of 429 elite white athletes from 14 different sporting disciplines and 436 controls, the sprint athlete group showed a higher frequency of the 577RR genotype (50%) and a lower frequency of the 577RX genotype (45%), compared with controls (30% and 52%, respectively), while the elite endurance athletes displayed a higher frequency of the 577XX genotype (24%) than controls (18%)55; however, the sample sizes in the truly elite subgroups are very small, and therefore, any conclusions drawn from them are prone to a high risk of type I error and should be treated with caution. Interestingly, MacArthur et al100 developed an exciting ACTN3 knockout (KO) mouse model in order to investigate the mechanisms underlying ACTN3 deficiency. These authors found that the KO mice had similar muscle fibre proportions as the wild type but reduced muscle mass, which appeared to be accounted for by the reduced fibre diameter of the fast-twitch muscle observed in KO mice.100 In addition to alterations in muscle fibre size, increased activity of muscle aerobic enzymes, longer muscle contracting time and shorter recovery period from fatigue were attributed to the characteristics of the ACTN3 KO mouse. Thus, the phenotypes of the ACTN3 KO mouse mimic the gene association studies performed in humans and provide a plausible explanation for the reduced sprint/power capacity and improved endurance performance in humans with the ACTN3 577XX genotype.

Implications of current genetic findings in association with elite athlete status

World-class athletic performance is a complex multifactorial phenotype, and it is acknowledged that to become an elite athlete, a synergy of physiological, behavioural and other environmental factors is required. It is commonly perceived that genetic endowment is one of the arbiters of elite athletic performance, a belief perhaps augmented by the striking geographical variation in athletic success.49 ,50 ,101 As reviewed above, a number of genes have been found to associate with elite performance. These studies have employed primarily the candidate gene approach to identify those genes which are associated either with elite performance or with variation in performance-related traits. As such, a number of genes have been found to associate with elite performance, although generally with small effect sizes and heavily prone to type I statistic error. The number of candidate genetic variants that can potentially explain elite athletic status will be much higher than those examined by numerous biotechnology companies such as Sports X Factor (eg, 7 genes: http://www.sportsxfactor.com). While genetic testing will very likely become a part of talent ID programmes in the future, current genetic testing is of almost zero predictive capacity despite testimonials to the contrary (www.xrgenomics.com/testimonials). Whether new approaches such as GWAS will significantly improve prediction outcomes for athletes is unknown.

The future

Most of the knowledge in sports genetics (including most of the information presented in this review) has been generated primarily using classical/old-fashioned genetic methods such as candidate gene analysis and almost exclusively applied to cohorts with small sample sizes (usually n≤300) and rather unsophisticated multifactorial phenotypes (eg, aerobic capacity and athlete status). The data thus generated from these studies and reviewed here need to be examined in the light of the view held by most ‘hard-core’ geneticists that a study of any complex phenotype in humans is futile unless a cohort size of between 20 000 and 100 000 is used and therefore possessing sufficient statistical power for meaningful analysis and interpretation. If one accepts this view (not currently held by the authors of this review), then all the studies reviewed here should be ignored. While this view is somewhat extreme, an intermediate view (currently held by the authors of this review) is that, perhaps beside the ACTN3 R577X and possibly ACE I/D, the vast majority of the candidate genes for sporting performance discovered to dateare not the key candidates seriously implicated in the phenotypes of interest. Priority should therefore be given to recruiting sufficiently large study cohorts with adequately measured phenotypes to increase statistical power. Some of the elite athlete cohorts described in this review may suffice, and collectively, these cohorts could be used for replication purposes.

As stated previously, it is accepted that there will be many genes involved in sporting performance, and hence it is timely that genetic research has moved to the genomics era. New approaches and technologies will no doubt be increasingly applied to searching the whole human genome instead of studying single genes or indeed SNPs as the cost of using such whole-genome methods becomes more affordable. In particular, the cost of large-scale sequencing has dramatically dropped, from the first complete human genome costing $3 billion to sequence in 2000 to $1000 per genome, as promised by the company Ion Torrent (a division of Life Technologies) using its new Ion Proton sequencer in 2012.102 At present, no matter the success or failure of the GWAS approach, this approach is certainly providing, and will continue to provide, the insight into genetic architecture and the molecular basis underlying human diseases and complex traits. The first genome-wide association study in age-related macular degeneration (AMD) revealed an intronic and common variant significantly related to AMD by comparing 96 cases with 50 controls, and consequently a functional polymorphism in the Complement Factor H (CFH) gene was identified by resequencing,22 suggesting that it is not unreasonable to expect and detect variants with large effects in a small study; however, large cohorts (ie, ranging from several thousands to 20 000–100 000) will be routinely studied by GWAS and will provide good resources for all scientific fields including genomics of the world-class athlete.103 This development will require a move away from the traditional way of researching in exercise science/medicine (ie, predominantly single-laboratory studies) to large, well-funded collaborations/consortia with leading industry partners and therefore substantial statistical/technological power and know-how. Only with such resources can the most strongly acting genes be identified with confidence, enabling gene×gene interactions to be revealed and gene×environment interactions to be studied more accurately. One first example in the area of sports performance has recently been successfully piloted.104 The overall purpose was to identify new SNPs that confer susceptibility to sprint and endurance performance by the use of world-class athletes as participants. GWAS was initially performed using the Illumina HumanOmni1-Quad BeadChip (>1 000 000 SNPs/sample) or the HumanOmniExpress BeadChip (>700 000SNPs/sample) in 95 sprinters and 102 controls from Jamaica. After removing individuals and markers failing quality control, the remaining SNPs were taken forward for association analysis. The genotype frequencies of 88 Jamaican sprinters and 87 Jamaican controls were compared using logistic regression (corrected for population stratification), assuming an additive model. Seventeen SNPs crossed a predetermined significance threshold of 5×10–5.40 Further validation of these signals in independent cohorts is underway, and the replicated SNPs will be taken forward for fine-mapping and functional studies to uncover the underlying biological mechanisms. Further analyses using other cohorts, such as the ones described in this review, will also provide opportunities for verifying these GWAS findings across different ethnicities.

What are the new findings?

  • Research has shifted from twin-based/family-based studies to the study of single nucleotide polymorphisms (SNPs) in populations.

  • Over 200 SNPs associated with physical performance have been reported.

  • Historically, genomics research has been hampered by small sample sizes.

  • Candidate gene analysis is the most commonly used approach, and it is effective in detecting genetic variants with a small or modest influence.

  • Genome-wide association approaches examine the association of genetic variation across a large number of SNPs simultaneously.

  • The number of large genetic cohorts of world-class athletes is limited.

  • With larger samples, genome-wide association would allow the detection of smaller gene effects and maximise the amount of variation captured.

  • Current genetic testing has zero predictive power on talent identification and should not be used by athletes, coaches or parents.

How might it impact on clinical practice?

  • Genetic markers of sports injuries (eg, tendinopathy) are being discovered and may be used in the future in conjunction with other health indices to provide personalised care for an athlete.

  • Genetic variants influencing elite athletic performance are also expected to impact on cardiac, skeletal muscle and energy metabolism.

  • Genomic data may eventually help to prevent, diagnose and treat diseases such as myocardial dysfunction and muscular-skeletal diseases, and to determine whether or not to tailor prevention and treatments to specific populations.

References

View Abstract

Footnotes

  • Competing interests None.

  • Provenance and peer review Commissioned; internally peer reviewed.

  • ▸ References to this paper are available online at http://bjsm.bmj.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles

  • Correction
    BMJ Publishing Group Ltd and British Association of Sport and Exercise Medicine