SOMBRERO v.1.1 is now available for download. See
below for a summary of changes in this update
of the software. The software is free to use, on condition that
any use of the software in a publication will be acknowledged by
citation of the relevant publications. Note that this software is
still very much under construction, so please contact the author
with any bugs or problems (shaun.mahony(AT)nuigalway.ie).
Binaries: Precompiled binaries
for various systems are available below. If your system does not
appear in the list, try compiling the source
yourself or otherwise contact the author.
Backgrounds: The "BackExtract.pl"
perl script included in the SOMBRERO release allows the user to
create their own background model for any organism. The following
frequently requested background files are provided for the user's
backgrounds.zip: This zip
file contains 3rd-order Markov models of intergenic DNA for the
genomes of E. coli, S. cerevisae, Drosophila
and Human. The human background may be suitable for other mammalian
genomes, but if in doubt, generate your own background.
Priors: Pre-trained binding profile
SOMs for use with the "-prior" option in SOMBRERO. At
the moment, users cannot train their own binding profile SOMs, and
only the list below is supported. All suggestions of other sources
for binding matrix datasets would be greatly appreciated.
SOMBRERO is released under the GNU
General Public Licence. The source code for SOMBRERO is available
from this link,
but no guarantees are offered as to the code's stability or even
if the code will compile on the user's system. Note that the "ftime"
timing function is deprecated on some systems. If you are getting
errors related to ftime, please use this
version of the code instead.
Source: A parallelised version of SOMBRERO has been written
using MPI. The source code is available from this
link, but again, we cannot guarantee whether or not the code
will compile/run on the user's system. This version of SOMBRERO
must be run using "mpirun", e.g. "mpirun -np 4 ./SOMBRERO"
or sometimes "mpirun -np 4 -machinefile $PBS_NODEFILE ./SOMBRERO",
where -np specifies the number of processors to use. Please talk
with your system's administrator about using an MPI program on the
Changes in version
- Support for including information on known transcription factor
binding matrices in the initialisation.
- Support for masking lowercase characters and N's.
- The size of the SOM is by default chosen automatically according
to the size of the input dataset.
- Complexity threshold can now be adjusted.
- Entirely random initialisation now allowed.
- Error in the SOMBRERO Viewer fixed.
- Other minor bugs fixed.