Running MrBayes

1
Running MrBayes

• Use execute to bring data in a Nexus file into
  MrBayes
• Set the model and priors using lset and prset
• Run the chain using mcmc
• Summarize the parameter samples using sump
• Summarize the tree samples using sumt
• Note that MrBayes runs two independent
  Metropolis-coupled analyses by default

2
Convergence Diagnostics

• By default, MrBayes performs two independent
  analyses starting from different random trees
  (mcmc nruns2)
• Average standard deviation of clade frequencies
  calculated and presented during the run (mcmc
  mcmcdiagnyes diagnfreq1000) and written to file
  (.mcmc). Suggested target lt 0.05.
• Standard deviation of each clade frequency and
  potential scale reduction for branch lengths
  calculated with sumt
• Potential scale reduction calculated for all
  substitution model parameters with sump
• Other tools Awty and Tracer

3
Some Good Advice

• Run at least 1,000,000 generations
• Dont worry if average standard deviation of
  split frequencies fluctuate, especially in the
  beginning of the run
• Save time by running the analysis without heating
  - works well for many analyses
• Use MPI version on multiprocessor Macs
• Save time by starting with randomly perturbed,
  good trees
• If it is easy to get topological convergence,
  improve sampling by decreasing frequency of
  topology proposals
• If you use heated chains and there are few swaps
  between chains, try to lower the temperature
  coefficient
• If you have difficulties with convergence
• Increase the number of heated chains
• Run the analysis longer
• Make the model more realistic
• Start with randomly perturbed good trees

4
Resources

• My page (www.scs.fsu.edu/ronquist/mrbayes)
• MrBayes web site (www.mrbayes.net)
• MrWiki manual (on the MrBayes web site)
• MrBayes 3.1 command reference
• MrBayes 3.1 manual (pdf) with two tutorials a
  simple analysis (primates.nex) and an analysis of
  partitioned data (cynmix.nex)
• Last pages of the manual contain graphical
  summaries of the MrBayes 3 models and most common
  types of proposal mechanisms
• MrBayes 3 chapter complex partitioned analysis
  (kim.nex)

5
(No Transcript)
6
(No Transcript)
7
MrBayes 3.2

• Check-pointing
• Better tree proposals
• Strict and relaxed clock models and dating
• Autotuning and manual setting of proposal tuning
  parameters
• Setting of starting values
• Gibbs sampling of discrete gamma
• Max or average standard deviation of split
  frequencies

8
Moves
Data Set
Time to convergence
9
Savolainen (2000, Syst. Bio.) 357 taxa 2,925 bp
atpB rbcL
Four runs, no heating
extending TBR
parsimony-biased SPR