Population structure and population history of Anopheles dirus mosquitoes in Southeast Asia.
Separating the confounding effects of long-term population history from gene flow can be difficult. Here, we address the question of what inferences about gene flow can be made from mitochondrial sequence data in three closely related species of mosquitoes, Anopheles dirus species A, C, and D, from southeast Asia. A total of 84 sequences of 923 bp of the mitochondrial cytochrome oxidase I gene were obtained from 14 populations in Thailand, Myanmar, and Bangladesh. The genealogy of sequences obtained from two populations of AN: dirus C indicates no contemporary gene flow between them. The F(ST) value of 0.421 therefore probably represents a recent common history, perhaps involving colonization events. Anopheles dirus A and D are parapatric, yet no differentiation was seen either within or between species. The starlike genealogy of their haplotypes, smooth unimodal mismatch distributions, and excess of low frequency mutations indicate population expansion in An. dirus A and D. This, rather than widespread gene flow, explains their low within-species F(ST) values (0.018 and 0.022). The greater genetic diversity of An. dirus D suggests that expansion occurred first in species D and subsequently in species A. The current geographical separation and low hybrid fitness of these species also argue against ongoing interspecific gene flow. They suggest instead either historical introgression of mtDNA from An. dirus D into species A followed by independent range expansions, or a selective sweep of mtDNA that originated in An. dirus D. While not excluding contemporary gene flow, historical population processes are sufficient to explain the data in An. dirus A and D. The genealogical relationships between haplotypes could not be used to make inferences of gene flow because of extensive homoplasy due to hypervariable sites and possibly also recombination. However, it is concluded that this approach, rather than the use of fixation indices, is required in the future to understand contemporary gene flow in these mosquitoes. The implications of these results for understanding gene flow in another important and comparable group of malaria vector mosquitoes in Africa, the An. gambiae complex, are also discussed.