The Beowulf Cluster at U.M.-St. Louis

• Home
• Using the Cluster
• Programs
  • Bayesian Dating
  • ClustalW
  • Lamarc
  • Migrate
  • Modeltest
  • MrBayes
  • PAML
  • PAUP
  • PAUP Ratchet
  • POY
  • QTL Cartographer
• Troubleshooting
• The Kellogg Lab
• UMSL Beowulf Site

PAUP on the cluster

PAUP version 4.0b10 is on the cluster.

You can run a single analysis using these instructions.

To run a bootstrap analysis:

1. Prepare your nexus data file. Make sure your file has Unix line breaks (use TextWrangler). Append to your file a PAUP block that contains the commands and options for running your analysis. Among other things, this block must include the Bootstrap command and the BSeed, NReps, and the Treefile options. NReps should be set to 1. The Treefile must have ".t" at the end. Don't worry about the number that you choose for the BSeed option, as this will be changed by one of the scripts. See the PAUP command line manual for information about these and other commands and options. Give your file a .nxs (not .nexus) extension.
   
   Here are two sample blocks:
   
   
   • For a likelihood bootstrap:
     
     BEGIN PAUP;
     set autoclose=yes warnreset=no increase=auto;
     set criterion=likelihood;
     Lset [your likelihood settings from Modeltest go here];
     Bootstrap BSeed=1 Nreps=1 Treefile=tree.t Search=heuristic/ addseq=random;
     END;
     
     For a parsimony bootstrap:
     
     BEGIN PAUP;
     set autoclose=yes warnreset=no increase=auto [maxtrees=500];
     set criterion=parsimony;
     Bootstrap BSeed=1 Nreps=1 Treefile=tree.t Search=heuristic/ addseq=random;
     END;
     
     In the case of parsimony and if you have a large data set, you should consider including the maxtrees option (e.g., Maxtrees=500) by removing the square brackets that surround this option and by changing "increase=auto" to "increase=no".
     
     NOTE: You should understand the commands and options, take a look at the PAUP command line manual. You may need to specify other options for your specific needs.
   • 
     
2. Using sftp (use Fugu), upload your nexus file to your directory on the cluster, ideally to an empty folder.
3. 
   
4. Open Terminal and within the folder that contains your data file, type "prepboot.paup", and follow the instructions. You will be asked for the name of your input file (your nexus data file), a beginning bootstrap seed number (type any odd number), and the number of bootstrap replicates you want to perform. The script will tell you when it is finished. This script will set everything up for you to run your bootstrap analyses.
5. 
   
6. Jobs are run on the cluster using the same script that we have always used - the runjobs.paup script. Type "runjobs.paup X Y" replacing X with the beginning data set number (usually 1) and Y with the ending data set number (usually 100). For example, to run 100 bootstrap replicates you would type: "runjobs.paup 1 100". If for some reason, you are only running one replicate you would type "runjobs.paup 1 1". Running this script will send all of your jobs to the queue. Note that one individual can run only 32 jobs at a time. Any jobs submitted over 32 will sit in the queue on hold and will begin to run as your running jobs finish.
7. 
   
8. When your jobs are finished, type "mkbootree.paup". You will be asked for the beginning data set number (e.g., 1), the ending data set number (e.g., 100), and for a name to give your tree file. This script will collect all of the trees from your individual bootstrap replicates and place them in a single file with a taxon translation table. This file can be opened in PAUP. Download this file off of the cluster (using sftp). Open PAUP and execute your original data set, WITHOUT the appended PAUP block that contains commands and options for running the analysis. Then use Trees > Get Trees From File to select the tree file you just downloaded. You can make a consensus tree to get a tree with bootstrap values.

To run a single analysis:

1. Prepare your nexus data file. Make sure your file has Unix line breaks (use TextWrangler). Append to your file a PAUP block that contains the commands and options for running your analysis. Note that you cannot use the blocks provided above - you must remove the bootstrap commands and options.
2. 
   Two example blocks are provided below. Remember that you should familiarize yourself with the PAUP Command-line Reference Document so that you will understand what you are doing.
   
   For a parsimony analysis:
   
   BEGIN PAUP;
   set autoclose=yes warnreset=no increase=auto [maxtrees=500];
   set criterion=parsimony;
   hsearch nreps=100 addseq=random swap=tbr;
   savetrees file=mp.tre brlens;
   END;
   
   For a likelihood analysis:
   
   BEGIN PAUP;
   set autoclose=yes warnreset=no increase=auto;
   set criterion=likelihood;
   Lset [your model settings go here];
   hsearch nreps=10 addseq=random swap=tbr multrees=on;
   savetrees file=ml.tre brlens append=yes;
   END;
   
   
   
3. Using sftp (use Fugu for this), upload your nexus file to your directory on the cluster, ideally to an empty folder.
4. 
   
5. Open Terminal and change to the directory with your data file. Within this directory type: "run1job.paup". You will be asked to enter the name of your data file.
6. 
   
7. When the script finishes, an email will be sent to your UMSL email account (i.e., yourgatewayid@umsl.edu). You can read the standard screen output by viewing the .out file.