The power to detect disease associations with mitochondrial DNA haplogroups

Genetic variation of mitochondrial DNA (mtDNA) has been linked to a number of multifactorial diseases, but there is currently no tool available to predict the optimal size for these investigations. We used a simulation-based (Monte Carlo) permutation test to generate power curves for European mtDNA haplogroup studies, to derive a universal equation to enable power calculations for prospective studies across the globe, and to show that very large cohorts are required to reliably detect an association with complex human diseases. In some populations, geographical variation in haplogroup frequencies will prevent the reliable detection of subtle haplogroup associations with uncommon disorders.