Indiana University
  •  
  •  
  •  

Using MEME on Big Red at IU

MEME (Multiple Em for Motif Elicitation) discovers patterns in nucleotide and amino acid sequences. MEME is developed and maintained by University of California - San Diego (UCSD), San Diego Supercomputer Center (SDSC), The University of Queensland (Australia), and the National Biomedical Computation Resource (NBCR) at UCSD.

On Big Red at Indiana University, MEME is installed at:

/N/soft/whatami/meme-3.5.4/

For user documentation, see the MEME web site.

Run MEME using the memejob script, which is in your path by default. Use options with memejob just as you would with MEME. For a list of all possible options, run memejob without any command-line options.

MEME is a parallel application that efficiently uses multiple processes. Doubling the number of processes typically halves the amount of time that MEME runs. You can request up to 128 processes without difficulty, but usually you'll notice a trade-off between the number of processes you request and the time it takes for your job to start running. Also, requests for 128 processes may be delayed before they are fulfilled. To request multiple processes for MEME, use memejob with the -p option to specify the number of processes:

memejob -maxsize 400000 -p 32 -protein C7683d

The above example submits a batch job using 32 processes (8 nodes x 2 processors per node x 2 cores per processor) to search for protein motifs in file C7683d (-maxsize is an option to MEME). If you omit the -p option, memejob will submit a job that uses 4 processes.

By default, your job is allowed to run for 2 hours. To request more time, use the -wallhours option. For example, to run the above job for 36 hours, use:

memejob -maxsize 400000 -p 32 -protein C7683d -wallhours 36

When you run memejob, 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.

MEME will produce files with filenames similar to memejob.999999.0.out, where 999999 is the number of your job. Errors and debugging output are stored in a separate files with filenames similar to memejob.999999.err.

The serial (i.e., single-process) version of MEME is also available on Big Red. To use it, add the meme keyword to your softenv environment. To make MEME permanently available at the command prompt, use:

echo +meme >> ~/.soft resoft

Note: In the serial version of MEME, the -p option is disabled, because the script does not handle LoadLeveler on Linux. You will receive an error message if you use the -p option.

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.