Phylogenetic signal and the utility of 12S and 16S mtDNA in frog phylogeny

Genes selected for a phylogenetic study need to contain conserved information that reflects the phylogenetic history at the specific taxonomic level of interest. Mitochondrial ribosomal genes have been used for a wide range of phylogenetic questions in general and in anuran systematics in particular. We checked the plausibility of phylogenetic reconstructions in anurans that were built from commonly used 12S and 16S rRNA gene sequences. For up to 27 species arranged in taxon sets of graded inclusiveness, we inferred phylogenetic hypotheses based on different a priori decisions, i.e. choice of alignment method and alignment parameters, including/excluding variable sites, choice of reconstruction algorithm and models of evolution. Alignment methods and parameters, as well as taxon sampling all had notable effects on the results leading to a large number of conflicting topologies. Very few nodes were supported in all of the analyses. Data sets in which fast evolving and ambiguously aligned sites had been excluded performed worse than the complete data sets. There was moderate support for the monophyly of the Discoglossidae, Pelobatoidea, Pelobatidae and Pipidae. The clade Neobatrachia was robustly supported and the intrageneric relationships within Bombina and Discoglossus were well resolved indicating the usefulness of the genes for relatively recent phylogenetic events. Although 12S and 16S rRNA genes seem to carry some phylogenetic signal of deep (Mesozoic) splitting events the signal was not strong enough to resolve consistently the inter-relationships of major clades within the Anura under varied methods and parameter settings.