In the variant-by-variant approach the genotype frequencies of each variant are compared using the one-sided Fisher's exact test, and the significance level of the group is found by Dunn-Sidak correction [38] of the smallest p-value in the group. Note that the Dunn-Sidak correction is very similar to the Bonferroni correction, as the Bonferroni correction is an approximation of the Dunn-Sidak correction. Whereas the Bonferroni correction is slightly conservative for independent tests (such as the independent variants in the power simulations), the Dunn-Sidak correction has the benefit of being exact.
