SCHULZE, A*; CUTLER, E.B.; GIRIBET, G.; Smithsonian Marine Station; Harvard University; Harvard University: Molecular and Morphological Evolution in Sipunculan Worms

Sipuncula are not only a key group in protostome evolution but they are also used as bio-indicators for marine pollution and as model systems in physiological studies. A stable phylogenetic framework is necessary to interpret results of these studies in an evolutionary context. Here we present our findings on the intra-phyletic relationships of the Sipuncula. Our analyses are based on four gene regions as well as 58 morphological characters. We sequenced the complete gene for 18S rRNA and obtained partial sequences for 28S rRNA, Histone H3 and the mitochondrial cytochrome c oxidase subunit I. The datasets were analyzed separately and in combination. First, we analyzed the data under direct optimization with the program POY. Subsequently, we used the implied sequence alignment resulting from the POY analysis and performed an analysis with Bayesian statistics under mixed models for the different data partitions. For the stem regions of the 18S rRNA, we applied a doublet model that accounts for correlated substitutions in the complementary strands of the ribosomal molecule. Both analyses support the monophyly of the Sipuncula and most of the same clades within the Sipuncula. They also agree that a clade containing the genera Sipunculus and Xenosiphon is the sister group to all other Sipuncula. They differ with respect to the relationships among the remaining clades but whereas the deep nodes in the POY analysis generally show jackknife support below 50%, most of the deep nodes in the Bayesian analysis are supported by 100% posterior probability. In conclusion, the direct optimization approach has been extremely useful for handling sequences of unequal length and generating conservative phylogenetic hypotheses. The application of mixed models in the Bayesian analysis greatly improved the resolution in the deep nodes.