Simple script to perform multiple pairwise comparisons of > 1 genomes using MUMmer and MUMi. Works best with shorter genomes (< 5 M bases)
114 ~2.2 M base genomes on intel i7-4790 @3.6 GHz with 8 threads approx. 1 hour to produce 6555 MUMi values
See paper below for information about MUMmer and MUMi:
A genomic distance based on MUM indicates discontinuity between most bacterial species and genera. Deloger M, El Karoui M, Petit MA. J Bacteriol. 2009 Jan; 191(1):91-9. doi: 10.1128/JB.01202-08
Download and install MUMmer from: http://mummer.sourceforge.net/ (requires gcc and make - use something like Cygwin/MinGW if using Windows)
If not already installed, run:
pip install dill
pip install numpy
Run:
./setup.py install
import mumi
if __name__ == '__main__':
mumi.mummer = PATH_TO/mummer # if mummer is not on SYS_PATH
mumi.batch_run(directory='PATH_TO/DIRECTORY', ext='fna', num_threads=4, k=19)
mumi.single_run('PATH_TO/genome1', 'PATH_TO/genome2', name1='genome1', name2='genome2', k=19)
num_threads = number of cpus to use (only applies to batch_run)
ext = file extension to use within directory e.g. .fna, .fasta, .fas (only applies to batch_run)
k = minimum unique match (MUM) length for mummer. NOTE: k should be > 14 to avoid non-specific MUMs and longer processing time; k = 19 works best for most cases.
name1/name2 = return names (only applies to single_run; names for batch_run are simply input files)
IMPORTANT:
use
if __name__ == '__main__':
...
statement to prevent infinite process spawning!
Please report any bugs with (preferably) or w/o suggested fixes
Contact me at mb1511@bristol.ac.uk for any more information
give_mumi.py is directly translated from the included Perl script (give_mumi.pl) in the mummer
package, but optimised for speed and efficiency as it is easier to communicate with from the
main script. The result is slightly (but not hugely) different with 3 significant figures of
similarity most of the time. I have much less experience with Perl, so please feel free to point
out any errors in translation.
The __init__.py, multi.py and fasta.py are tools and utils from a larger project and have been
stripped of some code, but feel free to use what is there for your own endeavours :)