RNA SEQUENCE DATA IN PHYLOGENETIC RECONSTRUCTION: TESTING THE LIMITS OF ITS RESOLUTION

Abstract— 18S ribosomal RNA sequences from 11 echinoderms are analysed using parsimony to investigate phylogenetic relationships. Their estimated divergence limes range from less than 20 Ma to more than 550 Ma before present. Phylogenies based on 18S rRNA sequence data are compared with well-established morphological phylogenies to discover at what evolutionary distance the two approaches start to produce incongruent results. Three regions of the 18S rRNA molecule are analysed separately and together, and paired and unpaired sites are also treated separately and combined.
Results show that a parsimony analysis of sequence data produces reliable results only when taxa have diverged more recently than about 100 Ma. At greater evolutionary distances (up to 250 Ma), paired nucleotides produce more reliable results than unpaired, while paired and unpaired data combined produce intermediate results. All trees within about 1% of the most parsimonious solution ought to be accepted. Transversions give results almost as reliable as paired regions though there were relatively few informative sites. The relationships of echinoderm classes, which diverged 450–550 Ma ago, are unresolved by 18S rRNA data.