In an article on allele frequencies for CODIS loci in 41 different populations, Budowle et al. (1) reported no significant deviations from Hardy-Weinberg expectations (HWE) after making theBonferroni correction for multiple tests. However, examination of the data reported in (1) for two Native American populations, Navaho and Salishan, shows significant departures from HWE at three loci for each population. Budowle et al. appear to have relied on an incorrect and unsupportable application of the Bonferroni correction (2,3) to disregard these deviations. These loci should not be used when the product rule is employed to compute the frequency of multi-locus genotypes in these populations unless appropriate corrective actions are taken to account for the observed deviations from HWE.
Available at: http://works.bepress.com/travis_doom/10/