Note on running NHPhylobayes

Current version is 0.2.3

Paul Lewis has a nice description of how to run nhPhylobayes here.

Nexus file format:

nhPhylobayes is very picky about the Nexus file format:

Must have “Begin data;” “Matrix” “End” and bases must be capitals. No “X”s for ambiguous amino-acids change to “-“

To get it to read a -T fixed topology the tree needs to be Phylip (valid Phylip tree files may not be read correctly but if you run it through PAUP, all is fine). All names should be 10 letters long (no spaces) in both the tree and the data matrix.

Running an analysis:

For example:

  • Current working dir is /home/cymon/nhpb
  • Data file is named ‘mydata.nex’
  • Submission is automatic to the SGE queue
  • Run number is 1
  • Number of modes (process categories) of the CAT model -n 110
  • Number of cycles to complete 10000
  • Rate of sampling of cycles 10
  1. Running chain

    Start nhbp:

    /share/apps/nhPhylobayes/nhpb -d mydata.nex -f run1 -v -z 10000 -e 10 -m catbp -n 110 -y /share/apps/nhPhylobayes/ -Q -x /home/cymon/nhpb
    

    Stop nhbp:

    /share/apps/nhPhylobayes/nhpb -s run1
    

    Restart nhbp:

    /share/apps/nhPhylobayes/nhpb -d mydata.nex -f run1 -m catbp -y /share/apps/nhPhylobayes/ -x /home/cymon/nhpb -Q
    
  2. Correlations between runs:

    compchain -f run1 -g run2 -b XXXX -c 3
    
  3. Get graphs of log-stamp for burnin:

    monitor -f run1 -p
    
  4. Statistics for chain:

    moncorrel -f run1 -b XXXX
    
  5. Get concensus tree:

    readtopo -x XXXX -c run1