Choice of topology estimators in Bayesian phylogenetic analysis

Wheeler WC and Pickett KM (2008. Topology-Bayes versus clade-Bayesin phylogenetic analysis. Mol Biol Evol. 25:447–453.)discuss two ways of summarizing the posterior probability distributionof a Bayesian phylogenetic analysis, which they refer to as"topology-Bayes" and "clade-Bayes." They claim that the clade-Bayesapproach leads to problems such as "exaggerated clade support,inconsistently biased priors, and the impossibility of topologyhypothesis testing," which are not problems for the topology-Bayesapproach. However, their argument for topology-Bayes over clade-Bayesis based on errors in the interpretation of summary statisticsassociated with Bayesian phylogenetic analysis. Although thereis a well-documented difference between the maximum posteriorprobability topology and the majority-rule consensus topology(the established terms for topology-Bayes and clade-Bayes summaries,respectively), both have a place in phylogenetic analysis. Choiceof summarization strategy should be driven by choice of parametersthat need to be estimated versus those to be marginalized giventhe evolutionary questions being asked or hypotheses being tested.