Lipscomb Diana L, Farris James S, Källersjö Mari, Tehler Anders
Department of Biological Sciences, George Washington University, Washington, D.C, 20052, U.S.A.
Naturhistoriska riksmuseet, Molekylarsystematiska laboratoriet, Box 50007, S-104 05, Stockholm, Sweden.
Cladistics. 1998 Dec;14(4):303-338. doi: 10.1111/j.1096-0031.1998.tb00341.x.
Sequences of the small subunit (SSU) ribosomal RNA are considered useful for reconstructing the tree of life because this molecule is found in all organisms and is large enough not to have become saturated with multiple mutations. However, these data sets are large, difficult to align, and have extreme biases in base compositions which makes their phylogenetic signal ambiguous. Large ambiguous data sets may have many most-parsimonious trees, and finding them all may be impossible using convential phylogenetic methods. To examine the reliability of the number and relationships of eukaryotic kingdoms proposed by previous analyses of the SSU, we calculated trees from aligned sequences from eukaryotes in the Ribosomal Database Project using parsimony jackknifing which uses a resampling procedure to rapidly search large data sets for the branches that are strongly supported and eliminates poorly supported groups. Two separate analyses were carried out: an analysis in which all bases were equally weighted, and one in which transversions only were used. The parsimony jackknife procedure was able to efficiently find trees in which most major groups of eukaryotes were supported and in which some evolutionary hypotheses proposed by previous workers were tested. The relationships of these major groups to each other were largely unresolved, indicating that the SSU data, as represented in this database, is insufficient for answering questions about these deep branches. Interestingly, the analysis of transitions differs from the results of the entire data set, primarily being less resolved. This indicates that transversional mutations are important contributors to the resolved structure of the tree.
小亚基（SSU）核糖体RNA的序列被认为有助于重建生命树，因为这种分子存在于所有生物体中，且足够大，不会因多次突变而饱和。然而，这些数据集规模庞大，难以比对，并且碱基组成存在极端偏差，这使得它们的系统发育信号模糊不清。大型模糊数据集可能有许多最简约树，使用传统的系统发育方法可能无法找到所有这些树。为了检验先前对SSU分析所提出的真核生物界数量和关系的可靠性，我们使用简约自展法从核糖体数据库项目中真核生物的比对序列计算树，该方法使用重采样程序快速在大型数据集中搜索得到有力支持的分支，并剔除支持不足的类群。进行了两项独立分析：一项分析中所有碱基权重相等，另一项分析中仅使用颠换。简约自展程序能够有效地找到支持大多数真核生物主要类群的树，并检验了先前研究人员提出的一些进化假说。这些主要类群之间的关系在很大程度上尚未解决，这表明该数据库中所呈现的SSU数据不足以回答有关这些深层分支的问题。有趣的是，转换分析的结果与整个数据集的结果不同，主要是分辨率较低。这表明颠换突变是树的解析结构的重要贡献因素。
