Adachi J, Hasegawa M
Department of Statistical Science, Graduate University for Advanced Studies, Tokyo, Japan.
J Mol Evol. 1995 Jun;40(6):622-8. doi: 10.1007/BF00160510.
The internal branch lengths estimated by distance methods such as neighbor-joining are shown to be biased to be short when the evolutionary rate differs among sites. The variable-invariable model for site heterogeneity fits the amino acid sequence data encoded by the mitochondrial DNA from Hominoidea remarkably well. By assuming the orangutan separation to be 13 or 16 Myr old, a maximum-likelihood analysis estimates a young date of 3.6 +/- 0.6 or 4.4 +/- 0.7 Myr (+/- 1 SE) for the human/chimpanzee separation, and these estimates turn out to be robust against differences in the assumed model for amino acid substitutions. Although some uncertainties still exist in our estimates, this analysis suggests that humans separated from chimpanzees some 4-5 Myr ago.
当位点间进化速率不同时,通过邻接法等距离方法估计的内部支长会显示出偏向于较短的偏差。位点异质性的可变-不变模型能很好地拟合人猿超科线粒体DNA编码的氨基酸序列数据。通过假设猩猩的分化时间为1300万或1600万年前,最大似然分析估计人类/黑猩猩分化的年轻时间为360万±60万或440万±70万年前(±1个标准误),并且这些估计结果对于氨基酸替换的假设模型差异具有稳健性。尽管我们的估计仍存在一些不确定性,但该分析表明人类与黑猩猩在约400万至500万年前分化。