When one is not enough: introgression of mitochondrial DNA in Drosophila.

The specific purpose of this study is to investigate alternate processes that may have resulted in the formation of the two distinct mitochondrial haplotypes (maI and maII) of Drosophila mauritiana. The most plausible explanation for the observed data is that there has been introgression of Drosophila simulans mtDNA into D. mauritiana. More generally, this article addresses three important issues concerning resolution of species relationships. First, pooling data from distinct process partitions has the potential to obscure biologically informative patterns of substitution (Bull et al. 1993; Ballard et al. 1998). Second, hypothesizing a species phylogeny from a single linkage partition may be problematic (reviewed by Doyle 1992; Avise 1994). Third, phylogenetic hypotheses may be influenced by the inclusion of a single representative of a terminal taxon (Omland, Lanyon, and Fritz 1999). Multiple unlinked loci and multiple individuals from within species that exhibit extensive population subdivision should be included to maximize the potential to fully resolve species relationships. Slowinski and Page (1999) identified two levels of potential error when inferring species phylogenies from molecular sequence data. First, a gene tree for a set of molecular sequences may be incorrectly inferred if there is sufficient systematic or random error (Swofford et al. 1996). Second, even if a gene tree is correctly inferred, deep coalescence, retention of ancestral polymorphism, gene duplication, and horizontal gene transfer among different species can produce a gene tree that differs from the species tree (Doyle 1992; Avise 1994). This second source of error may be confounded if the nucleotides are linked such that there is a single gene tree, or genealogy. As an example, genes in organelle genomes such as metazoan mtDNA are historically linked with no recombination (but see Awadalla, Eyre-Walker, and Maynard Smith 1999; Eyre-Walker, Smith, and Maynard Smith 1999; Hagelberg et al. 1999). In this case, phylogenetic hypotheses from different mtDNA genes can only infer the mitochondrial genealogy, which may or may not be the same as the true species tree. The well-corroborated mitochondrial genealogy of the Drosophila melanogaster subgroup implies that D. mauritiana is paraphyletic relative to D. simulans and Drosophila sechellia (Solignac and Monnerot 1986; Satta and Takahata 1990; Ballard 2000). One explanation for this result is that the mtDNA genealogy does not reflect the species relationships. The phylogenetic relationships