Note to users. If you're seeing this message, it means that your browser cannot find this page's style/presentation instructions -- or possibly that you are using a browser that does not support current Web standards. Find out more about why this message is appearing, and what you can do to make your experience of our site the best it can be.
Phylogenetic MCMC Algorithms Are Misleading on Mixtures of Trees
Elchanan Mossel1 and
Eric Vigoda2
Markov chain Monte Carlo (MCMC) algorithms play a critical rolein the Bayesian approach to phylogenetic inference. We presenta theoretical analysis of the rate of convergence of many ofthe widely used Markov chains. For N characters generated froma uniform mixture of two trees, we prove that the Markov chainstake an exponentially long (in N) number of iterations to convergeto the posterior distribution. Nevertheless, the likelihoodplots for sample runs of the Markov chains deceivingly suggestthat the chains converge rapidly to a unique tree. Our resultsrely on novel mathematical understanding of the log-likelihoodfunction on the space of phylogenetic trees. The practical implicationsof our work are that Bayesian MCMC methods can be misleadingwhen the data are generated from a mixture of trees. Thus, incases of data containing potentially conflicting phylogeneticsignals, phylogenetic reconstruction should be performed separatelyon each signal.
1 Department of Statistics, University of California at Berkeley, Berkeley, CA 94720, USA. E-mail: mossel{at}stat.berkeley.edu 2 College of Computing, Georgia Institute of Technology, Atlanta, GA 30332, USA. E-mail: vigoda{at}cc.gatech.edu
The editors suggest the following Related Resources on Science sites:
In Science Magazine
TECHNICAL COMMENTS
Fredrik Ronquist, Bret Larget, John P. Huelsenbeck, Joseph B. Kadane, Donald Simon, and Paul van der Mark (21 April 2006) Science312 (5772), 367a.
[DOI: 10.1126/science.1123622] |Abstract »|Full Text »|PDF »
THIS ARTICLE HAS BEEN CITED BY OTHER ARTICLES:
Estimating Species Trees: Methods of Phylogenetic Analysis When There Is Incongruence across Genes.