One of the most useful parameters that can be derived from population genetic data is the average heterozygosity (H), which has traditionally been employed to contrast levels of genetic variation among populations and to compare genetic variation among different loci. In spite of the increasing application of DNA fingerprinting to natural populations and to the genetic identification of humans, explicit methods for estimation of H from DNA fingerprinting data have not been developed. Contributing to this omission is the inability to determine relatively important genetic information for multilocus fingerprinting probes, such as the number of loci, the number of alleles, and the distribution of these alleles into specific loci. We have derived explicit calculations for the expected average heterozygosity and a maximum value for this estimate. These estimates are based upon the DNA restriction pattern matrices that are typical for fingerprinting studies of humans and natural populations. For several empirical datasets from our laboratory, these estimates are shown to be relatively close to each other. This observation, and consideration of the effect of missing alleles and alternate numbers of loci, suggest that the average heterozygosity can be accurately approximated using phenotypic DNA fingerprint patterns because this parameter is relatively insensitive to the lack of certain genetic information.
