Document Detail

Phylogenetic MCMC algorithms are misleading on mixtures of trees.
MedLine Citation:
PMID:  16195459     Owner:  NLM     Status:  MEDLINE    
Markov chain Monte Carlo (MCMC) algorithms play a critical role in the Bayesian approach to phylogenetic inference. We present a theoretical analysis of the rate of convergence of many of the widely used Markov chains. For N characters generated from a uniform mixture of two trees, we prove that the Markov chains take an exponentially long (in N) number of iterations to converge to the posterior distribution. Nevertheless, the likelihood plots for sample runs of the Markov chains deceivingly suggest that the chains converge rapidly to a unique tree. Our results rely on novel mathematical understanding of the log-likelihood function on the space of phylogenetic trees. The practical implications of our work are that Bayesian MCMC methods can be misleading when the data are generated from a mixture of trees. Thus, in cases of data containing potentially conflicting phylogenetic signals, phylogenetic reconstruction should be performed separately on each signal.
Elchanan Mossel; Eric Vigoda
Related Documents :
8676739 - Approximate methods for estimating the pattern of nucleotide substitution and the varia...
15764559 - Frequentist properties of bayesian posterior probabilities of phylogenetic trees under ...
14711089 - Recent results in cophylogeny mapping.
11933039 - Prediction trees with soft nodes for binary outcomes.
12554449 - Is cladogenesis heritable?
22169069 - Sensitivity analysis of bi-layered ceramic dental restorations.
1391689 - Effects of afterhyperpolarization on neuronal firing.
12651449 - Modeling the influence of temperature on monthly gross primary productivity of sugar ma...
9346059 - Validity of self reported data on injury prevention behavior: lessons from observationa...
Publication Detail:
Type:  Journal Article; Research Support, U.S. Gov't, Non-P.H.S.    
Journal Detail:
Title:  Science (New York, N.Y.)     Volume:  309     ISSN:  1095-9203     ISO Abbreviation:  Science     Publication Date:  2005 Sep 
Date Detail:
Created Date:  2005-09-30     Completed Date:  2005-10-07     Revised Date:  2007-03-19    
Medline Journal Info:
Nlm Unique ID:  0404511     Medline TA:  Science     Country:  United States    
Other Details:
Languages:  eng     Pagination:  2207-9     Citation Subset:  IM    
Department of Statistics, University of California at Berkeley, Berkeley, CA 94720, USA.
Export Citation:
APA/MLA Format     Download EndNote     Download BibTex
MeSH Terms
Bayes Theorem*
Likelihood Functions
Markov Chains*
Monte Carlo Method*
Comment In:
Science. 2006 Apr 21;312(5772):367; author reply 367   [PMID:  16627724 ]

From MEDLINE®/PubMed®, a database of the U.S. National Library of Medicine

Previous Document:  Preindustrial to modern interdecadal variability in coral reef pH.
Next Document:  Dynamic paternity allocation as a function of male plumage color in barn swallows.