Bayesian Phylogenetics of Bryozoa

The phylum Bryozoa is the largest phylum of the lophophorate invertebrates. The number of extant species has been estimated at around 5600 (Todd, 2000), but is probably greater (Hayward and Ryland, 1999). The phylum comprises three classes: Gymnolaemata, Stenolaemata and Phylactolaemata. The affinity of the Phylactolaemata with the rest of the Bryozoa is uncertain, and is disputed (Mundy et al., 1981). The Gymnolaemata consists of more than 3000 species classified in two orders, the Ctenostomata and the Cheilostomata. The Ctenostomata is further subdivided into two suborders, the Stolonifera and the Carnosa. The Cheilostomata first appears in Upper Jurassic horizons, and has remained the dominant bryozoan group (Taylor and Larwood, 1990). Four suborders are presently recognised within the Cheilostomata, the Inovicellata, Malacostegina, Neocheilostomina and Scrupariina, and the Neocheilostomina encompasses two infraorders, the Ascophora and the Flustrina (Gordon, unpublished). The class Stenolaemata is believed to have originated in the Lower Ordovician (approximately 480 mya) with the majority of taxa belonging to four extinct orders. The order Cyclostomata remains the only extant order and includes five (Taylor, 2000) or six (Kluge, 1975) suborders: Articulata, Cancellata, Cerioporina, Isoporina, Rectangulata and Tubuliporina. Phylogenetic studies of the Bryozoa are limited, and controversial in their findings. Despite the fact that numerous skeletal morphological characters are readily available, both for fossil and extant species, only a few studies have employed computerised cladistic methodologies in phylogenetic studies. However, these studies have been criticised by Todd (2000), who, on the basis of skeletal and morphological data for fossil and extant species, suggested that the Ctenostomata were paraphyletic, with Stenolaemata (Cyclostomata) and Cheilostomata nesting within the Ctenostomata. Polyphyly of the cheilostomes based on skeletal morphology has also been suggested by Gordon (2000) and Voigt (1991). Until very recently only three molecular phylogenetics studies had been conducted ([Dick et al., 2000], [Hao et al., 2005] and [Hao et al., 2002]) and results are inconclusive. Using 16S rRNA, Dick et al., (2000), found the Ctenostomata and Cheilostomata to be paraphyletic, with the Cyclostomata polyphyletic, and both cheilostomes and cyclostomes embedded within the Ctenostomata. These results bore little resemblance to the more commonly accepted tree topologies, in which Cyclostomata and Cheilostomata are monophyletic (Todd, 2000). However, the suitability of the 16S rRNA gene for phylogenetic studies is limited by its potential to resolve divergences only as far back as the mid-Cretaceous, which may be insufficient given the palaeontological record of the Bryozoa (Dick et al., 2000). Hao et al. (2002) employed 18S rRNA in their analysis, but used a limited number (12) of sequences, giving inconclusive, results. Subsequently, it has been suggested, following phylogenetic reconstruction analysis, that some of the sequences used in that study (Membranipora sp., AF119081, and Lichenopora sp., AF119080) may have represented contaminants (Waeschenbach, 2003). Finally, Hao et al., 2005 J.S. Hao, C.X. Li, X.Y. Sun and Q. Yang, Phylogeny and divergence time estimation of cheilostome bryozoans based on mitochondrial 16S rRNA sequences, Chinese Science Bulletin 50 (2005), pp. 1205–1211. Full Text via CrossRefHao et al. (2005) re-evaluated cheilostomate phylogenetic relationships based on the 16S rRNA gene and also presented results that conflicted with those of the morphological studies. A very recent study on the molecular phylogeny of Bryozoa is now available (Fuchs et al., 2009) based on 18S rDNA, 28S rDNA and the mitochondrial CO1 gene and using 32 species. This suggests monophyly of bryozoan classes, but ambiguous results for the relationships amongst them. In the present study the 18S rRNA gene has been used in a study of Bryozoa phylogenetics. The use of this gene is hindered by the difficulty of aligning the sequences due to the presence of variable regions in the secondary structure. However, the importance of an accurate alignment of 18S rRNA for successful retention of homologous characters within the aligned sequences has been emphasised by many authors, see Xia et al. (2003) for review. In phylogenetic studies of Bryozoa, secondary structure has not been used during alignment ([Dick et al., 2000], [Hao et al., 2005] and [Fuchs et al., 2009]), or if used, variable regions were excluded (Hao et al., 2002), an approach which has been criticised (Xia et al., 2003). In the present study the 18S rRNA was aligned using a secondary structure model