Explaining large mitochondrial sequence differences within a population sample

Abstract

Mitochondrial DNA sequence is frequently used to infer species' boundaries, as divergence is relatively rapid when populations are reproductively isolated. However, the shared history of a non-recombining gene naturally leads to correlation of pairwise differences, resulting in mtDNA clusters that might be mistaken for evidence of multiple species. There are four distinct processes that can explain high levels of mtDNA sequence difference within a single sample. Here, we examine one case in detail as an exemplar to distinguish among competing hypotheses. Within our sample of tree wētā ( Hemideina crassidens ; Orthoptera), we found multiple mtDNA haplotypes for a protein-coding region ( cytb / ND1 ) that differed by a maximum of 7.9%. From sequencing the whole mitochondrial genome of two representative individuals, we found evidence of constraining selection. Heterozygotes were as common as expected under random mating at five nuclear loci. Morphological traits and nuclear markers did not resolve the mtDNA groupings of individuals. We concluded that the large differences found among our sample of mtDNA sequences were simply owing to a large population size over an extended period of time allowing an equilibrium between mutation and drift to retain a great deal of genetic diversity within a single species.