Estimation of population allele frequencies from small samples containing multiple generations

Abstract

Estimations of population genetic parameters like allele frequencies, heterozygosities, inbreeding coefficients and genetic distances rely on the assumption that all sampled genotypes come from a randomly interbreeding population or sub-population. Here we show that small cross-generational samples may severely affect estimates of allele frequencies, when a small number of progenies dominate the next generation or the sample. A new estimator of allele frequencies is developed for such cases when the kin structure of the focal sample is unknown and has to be assessed simultaneously. Using Monte Carlo simulations it was demonstrated that the new estimator delivered significant improvement over the conventional allele counting estimator.