Assessing the reliability of 5S rRNA sequence data for phylogenetic analysis in green plants

Abstract

Elimination of compensating substitutions in portions of 5S rRNA sequences that exhibit secondary structure and scoring of common gaps as homologous characters affects both the topology of phylogenetic estimates within green plants and measures of their reliability. Allowing for a transition/transversion bias, on the other hand, has little affect. Detailed analysis of these data results in phylogenetic estimates largely congruent with widely accepted hypotheses of relationships among green plants, but bootstrapping provides little statistical support for most of these groupings. The lack of statistical support is a direct result of both the small number of phylogenetically informative sites in the 5S rRNA molecule and the high rate of change at those positions that are free to vary. When considered in conjunction with earlier distance-based analyses of phylogeny that use 5S rRNA, these results demonstrate that phylogenetic estimates based only on analyses of 5S rRNA sequences must be viewed with considerable caution.