School of Biological Sciences, Washington State University, Pullman, WA 99164-4236, USA.
School of Biological Sciences, Washington State University, Pullman, WA 99164-4236, USA.
Mol Phylogenet Evol. 2010 Apr;55(1):1-17. doi: 10.1016/j.ympev.2009.09.028. Epub 2009 Sep 26.
This study (1) uses nearly complete rRNA-gene sequences from across Metazoa (197 taxa) to reconstruct animal phylogeny; (2) presents a highly annotated, manual alignment of these sequences with special reference to rRNA features including paired sites (http://purl.oclc.org/NET/rRNA/Metazoan_alignment) and (3) tests, after eliminating as few disruptive, rogue sequences as possible, if a likelihood framework can recover the main metazoan clades. We found that systematic elimination of approximately 6% of the sequences, including the divergent or unstably placed sequences of cephalopods, arrowworm, symphylan and pauropod myriapods, and of myzostomid and nemertodermatid worms, led to a tree that supported Ecdysozoa, Lophotrochozoa, Protostomia, and Bilateria. Deuterostomia, however, was never recovered, because the rRNA of urochordates goes (nonsignificantly) near the base of the Bilateria. Counterintuitively, when we modeled the evolution of the paired sites, phylogenetic resolution was not increased over traditional tree-building models that assume all sites in rRNA evolve independently. The rRNA genes of non-bilaterians contain a higher % AT than do those of most bilaterians. The rRNA genes of Acoela and Myzostomida were found to be secondarily shortened, AT-enriched, and highly modified, throwing some doubt on the location of these worms at the base of Bilateria in the rRNA tree--especially myzostomids, which other evidence suggests are annelids instead. Other findings are marsupial-with-placental mammals, arrowworms in Ecdysozoa (well supported here but contradicted by morphology), and Placozoa as sister to Cnidaria. Finally, despite the difficulties, the rRNA-gene trees are in strong concordance with trees derived from multiple protein-coding genes in supporting the new animal phylogeny.
本研究(1)利用来自后生动物界(197 个分类群)的近乎完整的 rRNA 基因序列来重建动物系统发育;(2)展示了这些序列的高度注释、手动比对,特别参考了 rRNA 特征,包括配对位点(http://purl.oclc.org/NET/rRNA/Metazoan_alignment);(3)在尽可能消除破坏性、异常序列的情况下,测试如果采用似然框架是否可以恢复主要后生动物类群。我们发现,系统消除约 6%的序列,包括头足类、箭虫、等足类和多足类、栉水母和纽形动物的发散或不稳定排列的序列,以及苔藓动物和须腕动物的序列,导致支持环节动物、冠轮动物、原口动物和两侧对称动物的树。然而,后口动物从未被恢复,因为尾索动物的 rRNA 位于两侧对称动物的基部附近(无显著意义)。具有反直觉性的是,当我们对配对位点的进化进行建模时,与假设 rRNA 中所有位点独立进化的传统树构建模型相比,系统发育分辨率并没有提高。非两侧对称动物的 rRNA 基因比大多数两侧对称动物的 rRNA 基因含有更高的%AT。后生动物的 rRNA 基因发现有二次缩短、富含 AT 和高度修饰,这使这些蠕虫在 rRNA 树中位于两侧对称动物基部的位置产生了一些疑问——特别是栉水母,其他证据表明它们是环节动物,而不是这样。其他发现是有袋类和胎盘类哺乳动物、环节动物中的箭虫(这里得到很好的支持,但与形态学相矛盾),以及扁盘动物与刺胞动物的姐妹关系。最后,尽管存在困难,rRNA 基因树与从多个蛋白质编码基因中得出的树在支持新的动物系统发育方面具有很强的一致性。