SABINE has moved to github.com/draeger-lab/SABINE.

Download

You can download the latest version of SABINE on this page. As we would like to know, who is using our program in order to inform you about new versions, we kindly ask you to register before downloading the software. The registration is optional and no prerequisite for downloading the tool. You can find instructions for installing and using SABINE in the documentation section. The terms of use are specified in the license of this software.

Type	Size	Version
SABINE	51 MB	1.2 (2013-05-22)

Requirements

• SABINE requires a Linux platform.
• SABINE needs the latest JAVA JRE release (version 6 or later). You can get Java from www.java.com.

Release Notes

New Features in SABINE version 1.2

• Updated training set: The support vector regression models which play a central role in the SABINE algorithm were updated and additional transcription factors from more recent versions of public databases were incorporated as training data.

New Features in SABINE version 1.1

• Superclass auto-detection: The superclass of a given transcription factor can now be auto-detected based on homology to transcription factors in the training set of SABINE. Alternatively, the tool TFpredict can be used for superclass prediction.
• Dynamic Best Match Threshold: The Best Match Threshold corresponds to a cutoff for the predicted PFM similarity of known PFMs to the unknown PFM of the input transcription factor. This cutoff is now by default chosen dynamically depending on the PFM similarities predicted for the best matches. If best matches with very high DNA-motif similarity to the input factor were found, only these factors are used for the PFM transfer. If a lower but still significant DNA-motif similarity was found, SABINE can still perform a prediction using a lower cutoff value. For this purpose, three different cutoff values were predefined and associated to different confidence levels (high, medium, and low).
• DNA-binding domain prediction: The DNA-binding domains of a transcription factor can now be predicted using an alignment-based approach. As this is still an exploratory feature, it is currently only implemented in the command-line interface of the stand-alone version of SABINE. A viable alternative method is provided by the tool TFpredict which can also be employed for the prediction of DNA-binding domains.
• Installation validator: A new function was implemented which allows to automatically validate a local installation of SABINE.

Registration

We would like to know who is using SABINE. We will use this information only to keep track of downloads and to inform you about updates. Apart from this, your personal data will not be used for any other purpose nor divulged to third parties.

Name:
Organization:
E-mail:
Feedback:  (Optional)

License

This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.

You should have received a copy of the GNU General Public License along with this program. If not, see http://www.gnu.org/licenses/gpl.html.

Included third-party software

• The Local Alignment Kernel by Hiroto Saigo et al. was integrated in SABINE as a sequence feature to measure the similarity of the DNA-binding domains.
  Website: http://sunflower.kuicr.kyoto-u.ac.jp/~hiroto/project/homology.html
  
  
• We included the Mismatch Kernel by Christina Leslie et al. as an additional feature which incorporates the domain similarity of transcription factors.
  Website: http://cbio.mskcc.org/leslielab/software/string-kernels
  
  
• To capture the structural similarity of TF binding domains we applied the PSIPRED secondary structure prediction method by David Jones et al.
  Website: http://bioinf.cs.ucl.ac.uk/psipred/
  
  
• We integrated the Motif statistic software suite (MoSta) by Utz Pape et al. to measure the similarity of PFMs.
  Website: http://mosta.molgen.mpg.de/
  
  
• In order to generate familial binding profiles by merging PFMs, we applied the tool STAMP by Shaun Mahony et al.
  Website: http://www.benoslab.pitt.edu/stamp/
  
  
• For training and evaluation of the support vector regression models we used the libSVM implementation provided by Chih-Chung Chang et al.
  Website: http://www.csie.ntu.edu.tw/~cjlin/libsvm/
  
  
• We computed pairwise sequence alignment scores and generated sequence logos using the open source framework BioJava by Richard Holland et al.
  Website: http://biojava.org/wiki/Main_Page

Integrated data

The basis of our supervised machine learning based approach to predicting DNA-binding specificities of TFs was the generation of a non-redundant training data set. We restricted the sources to databases providing experimentally validated DNA-binding specificity information in terms of DNA-binding sites, consensus sequences or PFMs. Besides intergrating large databases spanning the whole eucaryotic kingdom, we extracted data from diverse smaller databases whose content is specific to particular organisms. An overview of our data sources can be found here:

TRANSFAC:		http://www.biobase-international.com/pages/index.php?id=transfac
JASPAR:		http://jaspar.cgb.ki.se/
YEASTRACT:		http://www.yeastract.com/
SCPD:		http://rulai.cshl.edu/SCPD/
AGRIS:		http://arabidopsis.med.ohio-state.edu/
FLYREG:		http://www.flyreg.org/