Department of Evolutionary Biology, University of Vienna, A-1090, Vienna, Austria.
Central Research Laboratories, Natural History Museum of Vienna, A-1010, Vienna, Austria.
BMC Evol Biol. 2020 Nov 4;20(1):144. doi: 10.1186/s12862-020-01699-0.
Phylogenetic relationships among the myriapod subgroups Chilopoda, Diplopoda, Symphyla and Pauropoda are still not robustly resolved. The first phylogenomic study covering all subgroups resolved phylogenetic relationships congruently to morphological evidence but is in conflict with most previously published phylogenetic trees based on diverse molecular data. Outgroup choice and long-branch attraction effects were stated as possible explanations for these incongruencies. In this study, we addressed these issues by extending the myriapod and outgroup taxon sampling using transcriptome data.
We generated new transcriptome data of 42 panarthropod species, including all four myriapod subgroups and additional outgroup taxa. Our taxon sampling was complemented by published transcriptome and genome data resulting in a supermatrix covering 59 species. We compiled two data sets, the first with a full coverage of genes per species (292 single-copy protein-coding genes), the second with a less stringent coverage (988 genes). We inferred phylogenetic relationships among myriapods using different data types, tree inference, and quartet computation approaches. Our results unambiguously support monophyletic Mandibulata and Myriapoda. Our analyses clearly showed that there is strong signal for a single unrooted topology, but a sensitivity of the position of the internal root on the choice of outgroups. However, we observe strong evidence for a clade Pauropoda+Symphyla, as well as for a clade Chilopoda+Diplopoda.
Our best quartet topology is incongruent with current morphological phylogenies which were supported in another phylogenomic study. AU tests and quartet mapping reject the quartet topology congruent to trees inferred with morphological characters. Moreover, quartet mapping shows that confounding signal present in the data set is sufficient to explain the weak signal for the quartet topology derived from morphological characters. Although outgroup choice affects results, our study could narrow possible trees to derivatives of a single quartet topology. For highly disputed relationships, we propose to apply a series of tests (AU and quartet mapping), since results of such tests allow to narrow down possible relationships and to rule out confounding signal.
关于多足亚门的各亚目（倍足纲、唇足纲、综合纲和少足纲）之间的系统发育关系仍然没有得到很好的解决。第一项涵盖所有亚目的系统基因组研究与形态学证据一致地解决了系统发育关系，但与大多数先前基于不同分子数据的系统发育树相冲突。外群选择和长枝吸引效应被认为是这些不一致的可能解释。在这项研究中，我们通过使用转录组数据扩展多足动物和外群分类群的取样来解决这些问题。
我们生成了 42 种节肢动物的新转录组数据，包括所有四个多足亚目和其他外群分类群。我们的分类群取样通过已发表的转录组和基因组数据得到了补充，产生了一个涵盖 59 个物种的超级矩阵。我们编译了两个数据集，第一个数据集涵盖了每个物种的全部基因（292 个单拷贝蛋白质编码基因），第二个数据集的基因覆盖范围较窄（988 个基因）。我们使用不同的数据类型、树推断和四分体计算方法来推断多足动物的系统发育关系。我们的结果明确支持单系的有颚类和多足动物。我们的分析清楚地表明，有一个单一的无根拓扑的强烈信号，但内部根的位置对选择外群很敏感。然而，我们观察到多足纲+综合纲以及唇足纲+倍足纲的强烈证据。
我们最好的四分体拓扑与当前的形态系统发育不一致，另一个系统基因组研究支持这种拓扑。AU 检验和四分体映射拒绝与形态特征推断的树一致的四分体拓扑。此外，四分体映射表明，数据集中存在的混杂信号足以解释源自形态特征的四分体拓扑的微弱信号。尽管外群选择会影响结果，但我们的研究可以将可能的树缩小到单一四分体拓扑的衍生物。对于存在高度争议的关系，我们建议应用一系列测试（AU 和四分体映射），因为这些测试的结果可以缩小可能的关系并排除混杂信号。
