Genetic variation in natural populations: problem of electrophoretically cryptic alleles.

Electrophoretic studies have shown that the average frequency of heterozygous loci per individual is about 12% in Drosophila and other invertebrates and about 6% in vertebrates. It is estimated that only about two-thirds of all amino acid substitutions change net electric charge; hence, a large fraction of all genetic variation may be undetected by electrophoresis. Peptide mapping of 11 independent alleles coding for alcohol dehydrogenase in Drosophila melanogaster has uncovered one cryptic variant; thus, the frequency of electrophoretically cryptic variation is apparently low, about 9% in this sample. Nevertheless, with a simple model it is shown that this degree of cryptic variation, if it is typical of other loci, would substantially change our current estimates of genetic variation: the average heterozygosity would increase from about 12% to about 25% for invertebrates and from about 6% to 21% for vertebrates. A variety of techniques--including sequential electrophoresis and heat or urea denaturation--have been used by various investigators to detect electrophoretically cryptic variation. These techniques appear to be less effective than peptide mapping for detecting cryptic variation, but, like peptide mapping, they suggest that standard electrophoresis may detect most of the protein variation present in natural populations. The charge-state model of protein variation proposes that the "alleles" detected by electrophoresis are extremely diverse classes consisting of many electrophoretically cryptic alleles. The alcohol dehydrogenase peptide-mapping results are inconsistent with the charge-state model.