Towards an assessment of character interdependence in avian RNA phylogenetics: A general secondary structure model for the avian mitochondrial 16S rRNA

Abstract

RNA has been a major source of information for animal phylogenetics for 20 years (see early examples in Carmean et al., 1992; Hillis and Dixon, 1989). Avian phylogenetics has benefited from comparisons of RNA in studies of a wide range of taxa at varying hierarchical levels. These comparisons have focused particularly on mitochondrial rRNA (i.e., 12S and 16S) and tRNA genes (e.g., Houde et al., 1997; Pereira et al., 2007; Tavares et al., 2006). The power of RNA genes as sources of phylogenetic information derives from their heterogeneous rates of evolution in different structural regions, which have the potential to resolve both recent and early phylogenetic events. In eukaryotes, the base-pairing regions of rRNA, or stems, evolve more slowly than loops, which are nonbase-pairing regions (Smit et al., 2007). Base-pairing patterns in RNA reflect complex secondary structures that consist of adjacent and anti-parallel Watson–Crick strands and G–U base-pairs (Brown and Ellis, 2005). Evolution of paired bases in RNA stems occurs, in part, via compensatory nucleotide changes, i.e., substitutions that maintain or restore base pairing following a base change on a complementary strand (Dixon and Hillis, 1993). This structural constraint has important consequences for phylogenetics, because paired bases may evolve in a non-independent fashion, a violation of the assumption of site-independent evolution common to most phylogenetic methods (Dixon and Hillis, 1993).