Fig 1.
Phylogeny of Ardisia and allies inferred using maximum likelihood (ML) based on nuclear ITS sequences.
Lower half of the tree. Numbers at branches are ML bootstrap support (ML BP) values and Bayesian posterior probabilities (BI). ML bootstrap support values are indicated only if at least 50, and the Bayesian posterior probabilities are indicated only if at least 0.7. Alternative branching between the two different analyses is showed as no support at the branch. Non-Ardisia species nested within the Ardisia generic complex are shown in boldface. Arrows indicate the presumed misidentification. Collection numbers (or other voucher information) are labeled after species names.
Fig 2.
Phylogeny of Ardisia and allies inferred using maximum likelihood (ML) based on nuclear ITS sequences.
Upper half of the tree. Numbers at branches are ML bootstrap support (ML BP) values and Bayesian posterior probabilities (BI). ML bootstrap support values are indicated only if at least 50, and the Bayesian posterior probabilities are indicated only if at least 0.7. Alternative branching between the two different analyses is showed as no support at the branch. Non-Ardisia species nested within the Ardisia generic complex are shown in boldface. Arrows indicate the presumed misidentification. Collection numbers (or other voucher information) are labeled after species names.
Fig 3.
Phylogeny of Ardisia and allies inferred using maximum likelihood (ML) based on the combined dataset.
Lower half of the tree. Combined dataset including nuclear ITS and two plastid intergenic spacer (psbA-trnH and rpl32-trnL) sequences. Numbers at branches are ML bootstrap support values and Bayesian posterior probabilities. ML bootstrap support values are indicated only if at least 50, and the Bayesian posterior probabilities are indicated only if at least 0.7. Alternative branching between the two different analyses was showed as no support at the branch. Non-Ardisia species nested within the Ardisia generic complex are shown in boldface. Arrows indicate the presumed misidentification. Collection numbers are labeled after species names.
Fig 4.
Phylogeny of Ardisia and allies inferred using maximum likelihood (ML) based on the combined dataset.
Upper half of the tree. Combined dataset including nuclear ITS and two plastid intergenic spacer (psbA-trnH and rpl32-trnL) sequences. Numbers at branches are ML bootstrap support values and Bayesian posterior probabilities. ML bootstrap support values are indicated only if at least 50, and the Bayesian posterior probabilities are indicated only if at least 0.7. Alternative branching between the two different analyses was showed as no support at the branch. Non-Ardisia species nested within the Ardisia generic complex are shown in boldface. Arrows indicate the presumed misidentification. Collection numbers are labeled after species names.
Fig 5.
Tanglegram for Ardisia and their symbiotic bacteria.
Symbiont species or clades (subtrees) are connected to their host by a gray line. Maximum likelihood trees based on the nrITS and plastid sequences (psbA-trnH and rpl32-trnL) for the host (left) and the 16S-23S rRNA of their symbiotic bacteria (right). Red box indicates monophyletic clade, and blue box indicates non-monophyletic clade. Numbers in the left tree are ML bootstrap support values. Collection numbers were labeled after species names.
Fig 6.
Distribution of costs of the reconstructions of the phylogenies.
Null distribution of costs of the reconstructions of the independent phylogeny onto the dependent phylogeny using random tip mapping (RTM) and random parasite trees (RPT) for randomization are shown with slash pattern bars and black bars, respectively. Observed distribution of costs of the reconstruction of the bootstrapping-dependent phylogenies is shown with white bars. Observed cost (= 75) of the reconstruction of the consensus-independent phylogeny onto the consensus-dependent phylogeny is shown with a vertical dotted line.