Cao Y, Adachi J, Janke A, Pääbo S, Hasegawa M
Institute of Statistical Mathematics, Tokyo, Japan.
J Mol Evol. 1994 Nov;39(5):519-27. doi: 10.1007/BF00173421.
The phylogenetic relationships among Primates (human), Artiodactyla (cow), Cetacea (whale), Carnivora (seal), and Rodentia (mouse and rat) were estimated from the inferred amino acid sequences of the mitochondrial genomes using Marsupialia (opossum), Aves (chicken), and Amphibia (Xenopus) as an outgroup. The overall evidence of the maximum likelihood analysis suggests that Rodentia is an outgroup to the other four eutherian orders and that Cetacea and Artiodactyla form a clade with Carnivora as a sister taxon irrespective of the assumed model for amino acid substitutions. Although there remains an uncertainty concerning the relation among Artiodactyla, Cetacea, and Carnivora, the existence of a clade formed by these three orders and the outgroup status of Rodentia to the other eutherian orders seems to be firmly established. However, analyses of individual genes do not necessarily conform to this conclusion, and some of the genes reject the putatively correct tree with nearly 5% significance. Although this discrepancy can be due to convergent or parallel evolution in the specific genes, it was pointed out that, even without a particular reason, such a discrepancy can occur in 5% of the cases if the branching among the orders in question occurred within a short period. Due to uncertainty about the assumed model underlying the phylogenetic inference, this can occur even more frequently. This demonstrates the importance of analyzing enough sequences to avoid the danger of concluding an erroneous tree.
利用有袋类动物(负鼠)、鸟类(鸡)和两栖类动物(非洲爪蟾)作为外类群,根据推断出的线粒体基因组氨基酸序列,估算了灵长目(人类)、偶蹄目(牛)、鲸目(鲸)、食肉目(海豹)和啮齿目(小鼠和大鼠)之间的系统发育关系。最大似然分析的总体证据表明,啮齿目是其他四个真兽亚纲目的外类群,并且无论氨基酸替换的假定模型如何,鲸目和偶蹄目与食肉目形成一个分支,食肉目是其姐妹分类单元。尽管偶蹄目、鲸目和食肉目之间的关系仍存在不确定性,但由这三个目形成的分支以及啮齿目相对于其他真兽亚纲目的外类群地位似乎已得到确凿证实。然而,对单个基因的分析并不一定符合这一结论,一些基因以近5%的显著性拒绝了假定正确的树。虽然这种差异可能是由于特定基因中的趋同或平行进化造成的,但有人指出,即使没有特殊原因,如果相关目之间的分支发生在短时间内,这种差异在5%的情况下也可能出现。由于系统发育推断所依据的假定模型存在不确定性,这种情况甚至可能更频繁地发生。这证明了分析足够多序列以避免得出错误树的风险的重要性。