Skip to main content
. 2007 Jan 12;73(5):1576–1585. doi: 10.1128/AEM.01996-06

TABLE 2.

Pairwise comparisons of the phylogenetic trees evaluated in this studya

Phylogenetic tree type Phylogenetic tree type
NJ MP1 MP2 ARB RAxML RAxML pars MrBayes
NJ 142.1 143.0 310.6 198.0 1024.3 333.9
MP1 0.60 17.5 252.7 155.6 1125.6 248.6
MP2 0.62 0.10 246.0 147.8 1125.5 246.3
ARB 0.71 0.70 0.67 287.8 1266.2 316.0
RAxML 0.67 0.60 0.59 0.75 1167.5 189.9
RAxML pars 0.68 0.57 0.56 0.73 0.59 1320.5
MrBayes 0.71 0.46 0.44 0.73 0.61 0.67
a

Shown are pairwise comparisons of the phylogenetic trees evaluated in Fig. 2 using the NDA (boldface type; upper triangle of matrix; results are arbitrary values with greater values indicating greater dissimilarity) and a partition metric (lower triangle of matrix; results are fractions indicating dissimilarity, ranging from 0 to 1). Methods are as follows: NJ, neighbor joining, as implemented in NEIGHBOR; MP1 and MP2, maximum parsimony, as implemented in DNAPARS; ARB, parsimony insertion, as implemented in Arb; RAxML, maximum likelihood, as implemented in RAxML; RAxML pars, RAxML parsimony insertion guide tree; MrBayes, Bayesian tree as implemented in MrBayes. All values along the diagonal are 0 for both methods (because each tree is identical to itself).