系统发生（Phylogeny）.docVIP

系统发生（Phylogeny） Evaluate phylogenetic trees and data There are already some programs that can evaluate phylogenetic signals and robustness of phylogenetic trees in the data (Swofford, et, al., 1997). For the former, the most popular method is compared with experimental data and random signal data (deflection and arrangement; experiment) for the latter, the observed data re sampling experiments support the evolutionary tree (non parametric bootstrap and doubling method). Likelihood ratio experiments can evaluate both the substitution model and the evolutionary tree. Random evolution tree (skew experiment) Simulation results show that the distribution of the random data sets generated by the random MP phylogenetic tree length is symmetrical, but the use of phylogenetic signal data sets, the distribution is asymmetric (Figure 9.9; Hillis and Huelsenbeck, 1992). In skew experiments, the critical values of G1 statistics vary with the number of taxa and the loci in the sequences. This experiment does not assess the reliability of a particular topology, and this experiment is sensitive to even the very few signals in other random data sets. If the data clearly support group, and group classification of these packets are selectively removed, so this experiment can be used to determine the phylogenetic signal is still preserved, of course, at least for the test provides 10 different signatures and 5 taxa. The program is included in PAUP. Random character data (permutation experiment) The random data method determines whether a MP phylogenetic tree obtained from real data or a part of it can be accidentally obtained. In fact, the data is not truly random, but in each comparison column arranged in different order, make the initial data is eliminated degeneration; resulting in a non random sequence of sequence alignment; correct to say that every one of these sites in the sequence are obtained from the random base group occupy the sites in the entire alignment in. At the end of