A CONGRUENCE TEST OF RELIABILITY USING LINKED MITOCHONDRIAL-DNA SEQUENCES

In the absence of certainty, well-corroborated hypotheses of species relationships serve as the best estimates of the true phylogenies of groups. This approach was extended to linked mitochondrial DNA (mtDNA) sequences that share the same gene phylogenies because of nonrecombination. This expectation of congruence forms the basis to test the reliability of unequal weighting for different base positions and changes of DNA sequences. The principal data for this test were the mitochondrial 12S ribosomal RNA (rRNA), 16S rRNA, and cytochrome c oxidase subunit II sequences of five pecoran species (Artiodactyla, Eutheria), as obtained from the same plasmid clones and/or mtDNA isolations of brain, liver, or heart. Heavy weighting for stems and first/second codon positions and for transversions were first evaluated against the molecular evolutionary properties of the three genes and then evaluated by congruence, as measured in terms of both agreement and resolution. With unequal weighting, congruence was increased among the trees that result from separate analyses of the three genes. These results support the reliability of these weighting schemes and illustrate how phylogenetic accuracy can be tested with gene phylogenies for linked sequences (e.g., of mtDNA).