**SiGN-BN** is
** gene network estimation software** using Bayesian network model and
nonparametric regression.
It can estimate regulatory dependencies between genes as

SiGN-BN implements several algorithms for estimating gene networks using Bayesian network model. Generally, because a Bayesian network requires huge computational time to learn its structure fitted to given gene expression data, it is not widely used for large-scale gene regulatory network analyses. Our research group develops several algorithms to overcome this problem using supercomputers. Currently, three algorithms are available depending on the size of gene networks: (a) the greedy hill-climbing (HC) algorithm + bootstrap method applicable to up to 1000 genes, (b) the neighbor node sampling & repeat (NNSR) algorithm applicable to genome-wide (whole genome) gene networks, and (c) parallel optimal structure search (Para-OS) algorithm for estimating the mathematically optimal network structures for small networks consisting of up to 37 genes.

**Currently, SiGN-BN is available for HGC Supercomputer and AICS K computer and/or its compatible systems.**

Updated: April 2 2019

- SiGN-BN HC+BS release 1.8.1 that supports Shirokane 5 (
`os7`environment) is now available under`~tamada/sign`on SHIROKANE. Script`signbn-hcbs.sh`automatically requests`os7`environment and the latest binary of SiGN-BN HC+BS. - The latest
`SiGN-Proc`is also available for`os7`environment. - SiGN-BN NNSR release 0.16.4 or later supports
`os7`environment. The updated script for the NNSR algorithm is also available as`~tamada/sign/signmpi.sh`. It automatically requests`os7`environment, therefore, you can execute the algorithm with new script and the same command line arguments. - Old SiGN-BN HC+BS binaries are able to run on Shirokane 4 with
`signbn-hcbs.os6.sh`script placed under`~tamada/sign`. This automatically requests`os6`environment (Shirokane 4) by "`-l os6`" option for Grid Engine in the script. - SiGN-Proc binary for
`os6`environment is available as "`signproc.os6`" under`~tamada/sign`. This works under`os6`environment. Therefore, if you want to use`signproc.os6`directly from the terminal, specify`-l os6`when you execute`qlogin`, e.g. "`qlogin -l os6`". If you use SiGN-Proc through Grid Engine, specify "`--bin signproc.os6`" for the job script. - SiGN-BN NNSR release 0.16.3 or earlier supports
`os6`environment. The job script is available as`~tamada/sign/signmpi.os6.sh`.

- Sep. 18, 2019 : SiGN-BN NNSR 0.16.6 Released.
- Sep. 12, 2019 : SiGN-BN NNSR 0.16.5 Released.
- Jul. 12, 2019 : SiGN-Proc 0.22.5 released.
- Apr. 2, 2019 : SiGN-BN NNSR 0.16.4 Released for Shirokane 5.
- Apr. 1, 2019 : SiGN-BN HC+BS 1.8.1 & SiGN-Proc 0.22.4 Released for Shirokane 5.
- Mar. 16, 2018 : SiGN-BN HC+BS 1.7.0 / NNSR 0.16.0 Released.
- Mar. 9, 2018 : SiGN-BN HC+BS 1.6.0 / NNSR 0.15.0 Released.
- Feb. 7, 2018 : SiGN-Proc 0.21.1 released.

**HC+Bootstrap:** Release 1.8.1

**NNSR:** Release 0.16.6

**Para-OS:** Release 0.1.2

How to use SiGN-BN : Online tutorial of SiGN-BN.

Manual : User reference manual.

Download : Download site for the binary distribution.

Contact : Contact information & developer list

Manual : User reference manual.

Download : Download site for the binary distribution.

Contact : Contact information & developer list

See SiGN Publication List for the complete list of
our publications using SiGN software. For publications specific to SiGN-BN algorithms,
see REFERENCE below.

SiGN-BN is developed in the ISLiM (Next-generation integrated simulation of living matter) project in RIKEN Computational Science Research Program. Computational resources required for the development of SiGN-BN was provided by the HGC Supercomputer System, Human Genome Center, Institute of Medical Science, The University of Tokyo; and RIKEN Supercomputer system RICC.

[1] Imoto, S., Goto, T., and Miyano, S. (2002).
Estimation of genetic networks and functional structures between genes
by using Bayesian network and nonparametric regression.
*Pacific Symposium on Biocomputing*, **7**, 175-186.
(PDF on
online proceedings)

[2] Tamada, Y., Shimamura, T., Yamaguchi, R., Imoto, S., Nagasaki, M.,
and Miyano, S. (2011).
SiGN: Large-scale gene network estimation environment for high performance computing,
*Genome Informatics*, **25** (1), 40-52.

[3] Tamada, Y., Imoto, S., Araki, H., Nagasaki, M., Print, C.,
Charnock-Jones, D.S., and Miyano, S. (2011).
Estimating genome-wide gene networks using nonparametric Bayesian
network models on massively parallel computers,
*IEEE/ACM Transactions on Computational Biology and Bioinformatics*, **8** (3), 683-697. (Pubmed)

[4] Tamada, Y., Imoto, S., and Miyano, S. (2011).
Parallel algorithm for learning optimal Bayesian network structure,
*Journal of Machine Learning Research*, **12**, 2437-2459.

[5] Honda, H., Tamada, Y., and Suda, R., (2016).
Efficient Parallel Algorithm for Optimal DAG Structure Search on Parallel Computer with Torus Network,
In Proceedings of *the 16th International Conference on Algorithms and Architectures for Parallel Processing* (ICA3PP 2016), Lecture Notes in Compute Sicence, **10048**, 483-502.

[6] Tamada, Y. (2018).
Memory Efficient Parallel Algorithm for Optimal DAG Structure Search using Direct Communication,
*Journal of Parallel and Distributed Computing*, **119**, 27-35, 2018.