Response to Comment on "Phylogenetic MCMC Algorithms Are Misleading on Mixtures of Trees"

doi:10.1126/science.1124180

Account Information

Guest Alerts | Access Rights | My Account | Sign In

Technical Comments

Response to Comment on "Phylogenetic MCMC Algorithms Are Misleading on Mixtures of Trees"

Elchanan Mossel¹^* and Eric Vigoda²^*

We presented a tree mixture in which Markov chain Monte Carlo(MCMC) methods have an exponentially slow convergence rate.We expect that many other mixture scenarios will show slow convergence.Ronquist et al. show that Metropolis-coupled MCMC (MC³) convergesquickly on our mixture. However, they presented no theoreticalor systematic experimental evidence determining the type ofmixtures where MC³ or other methods are efficient.

¹ Department of Statistics, University of California at Berkeley, Berkeley, CA 94720, USA.
² College of Computing, Georgia Institute of Technology, Atlanta, GA 30332, USA.

^* To whom correspondence should be addressed. E-mail: mossel{at}stat.berkeley.edu; E-mail: vigoda{at}cc.gatech.edu

Read the Full Text

The editors suggest the following Related Resources on Science sites:

In Science Magazine

TECHNICAL COMMENTS Comment on "Phylogenetic MCMC Algorithms Are Misleading on Mixtures of Trees": Fredrik Ronquist, Bret Larget, John P. Huelsenbeck, Joseph B. Kadane, Donald Simon, and Paul van der Mark (21 April 2006)
Science 312 (5772), 367a. [DOI: 10.1126/science.1123622]
| Abstract » | Full Text » | PDF »