Empirical estimates of Bonferroni corrections for use in chromosome mapping studies with the BXD recombinant inbred strains.

Most chromosome mapping efforts with the BXD recombinant inbred (RI) strains involve comparisons between a trait of interest and each of a large number of marker loci for evidence of linkage. Such multiple tests or comparisons greatly increase the Type I error rate compared to the single-test situation. Perhaps the most direct way to obtain multiple-test error rates is to employ a Bonferroni correction, where the single-test alpha is multiplied by the number of independent (nonredundant) comparisons (k) to yield a multiple-test alpha that protects against even one fortuitous association with any of the markers. Several empirical estimates of k are discussed for the published BXD marker set of 142 mapped loci and the newer unpublished marker set (October 1991) comprised of 352 marker loci. Reasonable estimates of k appear to be roughly 40 and 65 for these two marker sets, respectively.