The Closest Living Relatives of Land Plants Kenneth G. Karol, Richard M. McCourt, Matthew T. Cimino, and Charles F. Delwiche

View Appendix 2. (Note: This file has been revised from the originally posted version to add a number of references. The originally posted version is provided here.)

Appendix 3. Phylogenetic analyses.

The data were analyzed with two likelihood methods, Bayesian inference (BI) and maximum likelihood (ML), both of which are probabilistic methods that rely on explicit models of sequence evolution to test phylogenetic hypotheses. BI differs from ML in that it is based on the posterior distribution of a parameter conditional on the observations; the implementation of BI that we used relies on Markov Chain Monte Carlo methods to permit broad sampling of the data space. An important advantage of the Bayesian approach is that it is relatively fast and provides probabilistic measures of tree strength that are more directly comparable with traditional statistical measures than those more commonly used in phylogenetic analyses. Bayesian inference was carried out using MrBayes v2.0 (1). MrBayes uses a Metropolis-coupled Markov chain Monte Carlo (or MCMCMC) algorithm that runs several chains simultaneously. Two separate runs were carried out with four Markov chains, each starting from a random tree. Three of these chains were heated allowing for broad sampling of parameter space while the forth chain was not. The Markov chains were run for two million generations sampling every 100 generations for a total of 20,000 samples each run. The first 1000 samples from each run were discarded as burn-in (data points sampled before the chain reaches stationarity), and the remaining 38,000 samples (19,000 from each run) were combined into a single file and analyzed using the 'sumt' command in MrBayes. Both independent runs found essentially identical tree topologies and posterior probabilities, indicating that the sample number was sufficient to permit the algorithm to converge on a global solution.

Maximum likelihood analyses were carried out using PAUP* (2). Using the likelihood ratio test statistic, model selection procedures (3) identified the general-time-reversible model with invariant sites and gamma distributed rates for variable sites (GTR + I + ) as the best fitting model. With model parameters estimated from the data, a heuristic search with ten random taxon addition sequences, TBR branch swapping, and steepest descent option active was performed. Bootstrap analyses were conducted using 500 resampling replicates generated by CodonBootstrap (4) taking into account protein-coding and ribosomal portions of the dataset. For each bootstrap replicate, a heuristic search with three random taxon addition sequences, NNI branch swapping, and steepest descent option active was performed.

Analyses using maximum parsimony (MP), and minimum evolution with two distance measures [LogDet (ME-ld) and maximum likelihood (GTR+I+; ME-ml) distances] were also performed using PAUP* (2). For each analytical method, a heuristic search was performed with ten random taxon addition replicates, TBR branch swapping, and steepest descent option active. Bootstrap analyses were conducted using 1000 resampling replicates. Bootstrap values greater than 50% are not shown.

Appendix 4: Supplemental Figure 1. Phylogenetic relationships of the Charophyta inferred from the combined four-gene data set using additional analytical methods. Bootstrap values (1000 replicates) are shown above branches. An asterisk (*) denotes bootstrap values below 50%. (A) The maximum parsimony tree (tree score, 12,365 steps; scale bar, 100 substitutions). (B) Minimum evolution tree using LogDet distances (tree score, 2.50594; scale bar, 0.05 substitutions/site). (C) Minimum evolution tree using maximum likelihood distances (tree score, 3.57107; scale bar, 0.05 substitutions/site). **Bootstrap value for a very short branch uniting Chlorokybus with the remaining charophyte lineages. The topologies are drawn with Cyanophora rooting the trees. Taxonomy is modified from (23).

Sections A & B

Medium version | Full size version

Section C

Medium version | Full size version

References
1. J. P. Huelsenbeck, F. Ronquist, Bioinformatics 17, 754-755 (2001).
2. D. L. Swofford. PAUP* v4.0b8: Phylogenetic Analysis Using Parsimony * (and other Methods) (Sinauer, Sunderland, MA, 1998).
3. N. Goldman, J. Mol. Evol. 36, 182-198 (1993).
4. J. P. Bollback, CodonBootstrap v3.0b4 (Distributed by the author. Department of Biology, University of Rochester, Rochester, NY, 2001).