作者: Alan R. Templeton
DOI: 10.1111/J.1558-5646.1983.TB05533.X
关键词:
摘要: Recombinant DNA technology provides evolutionary biologists with another tool for making phylogenetic inference through contrasts of restriction endounuclease cleavage site maps or sequences between homologous segments found in different groups. This paper is limited to the problem from maps. Several methods such have already been used proposed (Avise et al., 1979a, 1979b;NeiandLi, 1979; Ferris 1981), but all these depend upon assumption that shared sites reflect common origins and are not result convergent evolution. Unfortunately, evolution occurs high probability this type data (Templeton, 1983). In addition, several enzymes generally pooled analyses. Recently, Adams Rothman (1982) examined distributions related 54 endonucleases. They concluded 1) distributed non-randomly most sequences, 2) there considerable heterogeneity (even those recognition same length) respect number distribution their respective 3) relationship based on distances will be biased. Brown al. sequenced 896 base pairs mitochondrial humans apes about 90% substitutions were transitions. The predominance transitions over transversions increases convergence expected when assumed equally likely Therefore, a need exists an algorithm deals more directly statistical inhomogeneity enzymes. paper, I propose algorithm. After discussing estimation tree, task testing then addressed. First, present non-parametric framework fit one hypothesized phylogeny versus alternative phylogeny. Second, procedures presented hypotheses relative rates among various lineages.