SiGN-BN Download
Latest Release
| SiGN-BN HC+BS: | 1.8.3 |
| SiGN-BN NNSR: | 0.16.8 |
| SiGN-BN Para-OS: | 0.1.2 |
| SiGN-Proc: | 0.22.6 |
See Change History for the history of the releases.
License
The SiGN-BN executable binaries are available under the following terms.
Use Restrictions
You may not:
- modify, translate, reverse engineer, decompile, disassemble or
otherwise attempt to reconstruct or discover the source code from
the binaries of the Software, except to the extent applicable laws
specifically prohibit such restriction.
- create derivative works based on the software
(e.g. incorporating the software in a commercial product or service
without a proper license).
- copy the software.
- rent, lease, sublicense, convey, distribute or otherwise
transfer rights to the software.
- remove any product identification, copyright, proprietary
notices or labels from the software.
Limitations of Liability
In no event shall the initial developers or copyright holders be
liable for any damages whatsoever, including - but not restricted to -
lost revenue or profits or other direct, indirect, special, incidental
or consequential damages, even if they have been advised of the
possibility of such damages, except to the extent invariable law, if
any, provides otherwise.
No Warranty
The software and this license document are provided AS IS with NO
WARRANTY OF ANY KIND, INCLUDING THE WARRANTY OF DESIGN,
MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE.
The SiGN-BN binaries include
SFMT
Copyright © 2006, 2007 Mutsuo Saito, Makoto Matsumoto and Hiroshima
University, & Copyright © 2012 Mutsuo Saito, Makoto Matsumoto, Hiroshima University
and The University of Tokyo.
Please cite our publications (see PUBLICATIONS in the home page)
when you publish your results.
Human Genome Center Shirokane 5
The latest binary and related files can be found in the author's home directory
at ~tamada/sign/.
You do not need to download the binary and install it by yourself.
General x86-64 Linux Systems
SiGN-BN HC+BS
- Release 1.8.3 binary for Linux x86-64 (signbn.1.8.3.gz)
- Gzipped executable binary compiled for Linux x86-64.
- The required libraries are statically linked. Therefore, this should work on every general
linux x86-64 system.
- Confirmed to work on CentOS 8 with AMD EPYC 7502P, CentOS 7 with Intel(R) Xeon(R) Gold 6154.
SiGN-BN NNSR
- Release 0.16.8 binary for Linux x86-64 (signbnnnsr.0.16.8.gz)
- Gzipped executable binary compiled for Linux x86-64.
- This is parallelized with Open MPI 4.1.0. You require
Open MPI installed on your system and to execute this via mpirun command.
- Set the Open MPI library path to the environment variable LD_LIBRARY_PATH if you have
an error about the missing library file.
- FYI: Using a server with Intel Xeon Platinum 8160 (2.1GHz) CPUs and 64 processes,
it takes 1 hour and a half to finish the network estimation with "-T 15000"
option for GN-10k-500.edf.txt, which can be
downloaded from below.
SiGN-Proc
- Release 0.21.1 binary for Linux x86-64 (signproc.0.21.1.gz)
- Gzipped binary compiled for Linux x86-64, and hopefully works on general intel 86-64 linux systems.
- Confirmed it works on the following environments.
- RIKEN AICS K computer login node.
- Oakleaf login node (Fujitsu FX-10 installed on Univ. Tokyo).
RIKEN R-CCS Fugaku computer and its compatible systems
SIGN-BN NNSR
- Release 0.16.7D (signbnnnsr.0.16.7D.gz) on Feb. 8, 2021.
- Compiled with lang environment "4.3.1 tcsds-1.2.29"
- Example of the job script: fugaku_job.sh.gz
- With the settings as in the above script and the sample data set below (GN-10k-500.edf.txt), Fugaku took less than 30 minutes to finished the network estimation (T=100,000).
- The degree of parallelism with a sufficient accuracy was confirmed up to 6144 processes (128 nodes × 48 processes/node, Flat MPI). Note: The number of processes affects the accuracy of the network estimation.
- This work used computational resource of Fugaku throught the HPCI System Research Project (Project ID: hp200194).
- Older release (Dec. 1, 2020): signbnnnsr.0.16.7C.gz compiled with lang environment "4.3.0a tcsds-1.2.28".
RIKEN AICS K COMPUTER and its compatible systems
SiGN-BN HC+Bootstrap
- Release 1.5.5 K-1.2.0-17 (signbn-hcbs.1.5.5_K-1.2.0-17.zip)
- Compiled with MPI+OpenMP hybrid parallelization.
- Compiled under the language environment K-1.2.0-17.
- This is not the open source software. DO NOT re-distribute the binary file.
SiGN-BN NNSR
- Release 0.14.2 K-1.2.0-17 (signbn-nnsr.0.14.2_K-1.2.0-17.zip)
- Compiled with MPI+OpenMP hybrid parallelization.
- Compiled under the language environment K-1.2.0-17.
- Set environment variable OMP_NUM_THREADS=8 in your job script.
- This is not the open source software. DO NOT re-distribute the binary file.
SiGN-BN Para-OS
- Release 0.1.2 K-1.2.0-21 (signbn-paraos.0.1.2_K-1.2.0-21.zip)
- Compiled with MPI+OpenMP hybrid parallelization.
- Compiled under the language environment K-1.2.0-21.
- Set environment variable OMP_NUM_THREADS=n in your job script where n represents the number of threads per CPU. Use n=8 for K computer.
- This is not the open source software. DO NOT re-distribute the binary file.
Sample Data
For SiGN-BN HC+BS
- GN-527-100.edf.txt.gz - 527 nodes (genes) x 100 samples in gzipped EDF format (222,051 Bytes).
- GN-527-500.edf.txt.gz - 527 nodes (genes) x 500 samples in gzipped EDF format (1,095,070 Bytes).
- GN-527-120-DY.edf.txt.gz - 527 nodes (genes) x 40 samples x 3 replicates for dynamic Bayesian networks in gzipped EDF format (232,855 Bytes).
- GN-527-24-DY.edf.txt.gz - 527 nodes (genes) x 8 samples x 3 replicates for dynamic Bayesian networks in gzipped EDF format (49,067 Bytes).
- GN-527.sgn3.gz - Gene network with 527 nodes (genes) used to generate above datasets
in gzipped SGN3 format (11,312 Bytes).
- GN-527-100-result.sgn3.gz - An example of the result of the network estimation by the bootstrap method with the above dataset GN-527-100.edf.txt.
For SiGN-BN NNSR
These files are the same as ones used in Tamada et al. (2011).
- GN-10k-500.edf.txt.gz - 10540 nodes (genes) x 500 samples in gzipped EDF format (21,979,935 Bytes)
- GN-10k-500-result.sgn3.gz - An example of the result of the network estimation by the NNSR algorithm with above dataset GN-10k-500.edf.txt.
- GN-10k-true.sgn3.gz - 10540 nodes (genes) gene network used to generate the above sample data set in gzipped SGN3 format (226,489 Bytes).
For SiGN-BN Para-OS