On Big Red, how do I submit batch jobs that run MrBayes?

MrBayes is a program for estimating phylogeny based on the posterior probability distribution of trees conditioned on a set of observations, using Bayes' theorem. MrBayes uses the Markov chain Monte Carlo simulation technique.

The parallel version of MrBayes is available on Big Red at:

For support documentation, see the MrBayes Help page.

On Big Red, you can use a script named mrbayesjob to submit a job that runs MrBayes. Your mrbayesjob script should be included in your path by default, and its manual page should be in the default path of manual pages. To run mrbayesjob, use:

Replace filename with the name of your MrBayes input file, n with the number of processors to use, h with the maximum amount of time that the job will be allowed to run, and queue_name with the name of the queue. (Only the filename is required.)

By default you get 4 processors for 2 hours on Big Red's LONG queue. To request more processors, specify an integer for -CPUS that is a multiple of 4. If you specify a value that is not a multiple of 4, the value will be increased to the next multiple of 4. The maximum number of processors for the LONG queue is 128.

For example, to use 16 processors to perform analysis specified in file mbdata in your current working directory, use:

To request a longer run time, specify an integer for -wallhours. The maximum run time is 336 hours (14 days) on the LONG queue.

Queues other than LONG are available; see Big Red usage policies.

When you run mrbayesjob, you'll receive a message when your job is submitted to the queue, and another when the job finishes. To check the status of your job, use the llq command.

In addition to output files that MrBayes produces, mrbayesjob will produce files with filenames similar to mrbayesjob99999.out and mrbayesjob99999.err, where 9999 is the number of your job. Such files contain information that MrBayes would print to the screen if you were running it interactively from the command line. They are the Unix standard output and standard error, respectively.

This document was developed with support from the National Science Foundation (NSF) under Grant No. 0503697 to the University of Chicago and subcontracted to Indiana University. Additional support was provided by IU through its participation in the TeraGrid, which is supported by the NSF under Grants No. 0833618, SCI451237, SCI535258, and SCI504075. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the NSF.

Comments/Questions/Corrections