Genome-wide significance is widely defined as having a statistical association in which the p-value ≤ 5 × 10^-8 in a GWAS.
As with all p-value thresholds for ‘statistical significance’, the value is arbitrary. The threshold of p ≤ 5 × 10-8 was a Bonferroni correction of the commonly used p ≤ 0.05, based on the estimated number of independent tests in the GWAS if all common SNPs in the first version of HapMap were tested with direct genotyping or imputation (i.e., 0.05/106). More stringent thresholds should be used if the study also included rare variants. The major feature of a reliable genetic association; however, is replication in an independent sample.
References
- Panagiotou OA, Ioannidis JP, Genome-Wide Significance P. What should the genome-wide significance threshold be? Empirical replication of borderline genetic associations. Int J Epidemiol 2012;41:273-86.
- Sterne JAC, Davey Smith G. Shifting the evidence - what's wrong with significance tests? BMJ (Clinical research ed) 2001;322:226-231.