No Evidence of a Common DNA Variant Profile Specific to World Class Endurance Athletes

1,520 endurance athletes (835 World Champs/Olympic) + 2,760 controls, 9 cohorts

Elite endurance athletes and ethnicity-matched controls: Australia, Ethiopia, Japan, Kenya, Poland, Russia, Spain, US

Two-stage case-control design: discovery in two cohorts, prespecified replication in seven. Largest elite-endurance genomic sample assembled; the null is informative but cannot exclude many small-effect alleles. The two discovery cohorts used different platforms and were run independently before GAMES existed (GENATHLETE on the Illumina CardioMetabochip, ~143,000 QC-passing SNPs, which is targeted cardiometabolic coverage rather than a full genome scan; the Japanese cohort on HumanOmniExpress, 541,179 autosomal SNPs). Discovery power was weak by design: 80% power to detect odds ratios of only 2.7/2.1 (GENATHLETE) and 6.1/4.5 (Japan) at MAF 0.1/0.3 and alpha 5x10-8. Replication genotyping was analysed centrally at Pennington. Control groups were not uniform — general-population or recreationally active in some cohorts, confirmed sedentary in others. Two internal inconsistencies to note: the paper says 1,045 (69%) of athletes were world-class in the text but 835 in the abstract and Table 1, and reports the discovery-excluded meta p-value for rs558129 as 0.0019 in Results and 0.0071 in Discussion. Competing interest: senior author C. Bouchard sat on the science advisory board of Pathway Genomics (2011-2014) — a direct-to-consumer genetic testing firm — though the reported null runs against that interest.

Discovery GWAS (GENATHLETE + Japanese runners) produced no genome-wide-significant hits at all (p < 5x10-8) in either cohort, and there was no overlap between the two cohorts' top SNPs; 45 markers at p < 1x10-4 were carried forward. In the prespecified replication cohorts, no marker survived multiple-testing correction — not in the four European cohorts, not in the two African cohorts (where the strongest Kenyan hits pointed in the OPPOSITE direction in Ethiopians). Conclusion stands: "no panel of genomic variants common to these elite endurance athlete groups", and "common genetic variants do not appear to be strong determinants of elite endurance athlete status". Heritable does not mean genotype-predictable: DNA-based talent identification is unsupported. Two corrections to the previous abstract-level record. (1) The null is not absolute — pooling all 8 available cohorts, ONE marker (rs558129 at GALNTL6) did survive Bonferroni correction (OR 0.81, p = 0.0002, I2 = 0, same direction in every cohort), and the paper's own summary reports it. It was only nominal (p = 0.04) when restricted to world-class athletes, and no other SNP survived. (2) The strongest result for the archive's purposes is not in the old record at all: the classic candidate genes failed outright. ACTN3, CKM and GNB3 showed no association in any cohort or meta-analysis, and none of 161 SNPs across 13 named endurance candidate genes (including ACE, ACTN3, NOS3, PPARA, PPARGC1A, ADRB2, AMPD1) reached the Bonferroni threshold of p = 3.5x10-7.

Molecular counterpart to pedigree/twin estimates: high familial h2 with no replicable loci, consistent with extreme polygenicity plus ethnic stratification.

"failed" is scoped to the prespecified replication phase: none of the 45 discovery leads survived multiple-testing correction in the seven replication cohorts, and directionally inconsistent nominal hits (rs7947391 across Australia/Poland/Spain; Kenya vs Ethiopia) are the signature of noise rather than of underpowered truth. This study is itself a failed replication of the prior candidate-gene literature (ACTN3, ACE, PPARGC1A and 10 other named genes). The single surviving pooled-meta marker (rs558129, GALNTL6) is a NEW discovery from the combined sample, not a replicated one, and has not to this record's knowledge been independently confirmed since.