Most current linkage analyses make use of highly polymorphic DNA markers. Assigning correct allele frequencies for these markers may be extremely difficult in particular study populations. Designation of erroneous frequencies may result in false-positive evidence for linkage, as well as in failure to correctly exclude linkage. These effects are most pronounced in small pedigrees with key individuals unavailable for typing. The power to correctly detect true linkage does not appear to be greatly affected by inaccurate allele frequencies. Before linkage analyses are performed for specific pedigrees, it is recommended that simulation analyses be performed, followed by uncertainty and sensitivity analyses.