Estimation of relative electrophoretic mutation rates from rare alleles in a sample

The theoretical justifications for using the number of rare alleles observed in a sample and the heterozygosity contributed by such alleles to estimate the relative electrophoretic mutation rate (REMR) are given in this note. It is shown that the estimator using the number of alleles has comparatively less bias. While the total heterozygosity contributed by all alleles at a locus has been previously used to estimate REMR with success, an analogous estimator with ony rare alleles has large bias over a wide range of effective population size and sample size.