HIERARCHICAL VERSUS NONHIERARCHICAL PATTERNS OF GENETIC DISTANCES AMONG POPULATIONS - A SIMULATION STUDY

Abstract

A simulation study was made of the effects of mixing two evolutionary forces (natural selection and random genetic drift), combined in a single data matrix of gene frequencies, on the resulting genetic distances among populations. Twenty-one, kinds of simulated gene frequencies surfaces, for 15 populations linearly distributed over geographic space, were used to construct 21 data matrices, combining different proportions of two types of surfaces (gradients and random surfaces). These matrices were analysed by Unweighted Pair-Group Method - Arithmetic Averages (UPGMA), clustering and Principal Coordinate Analysis. The results obtained show that ordination is more accurate than UPGMA in revealing the spatial patterns in the genetic distances, in comparison with results obtained using the Mantel test comparing directly genetic and geographic distances.