Training and using TFFM/PSSM/4-bits + DNA shape classifiers
Python Shell
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Failed to load latest commit information.


Python module for TFFM/PSSM/4-bits + DNA shape classifiers

This module allows for:

  1. Training PSSM/TFFM/4-bits + DNA shape classifiers on ChIP-seq data
  2. Applying PSSM/TFFM/4-bits + DNA shape classifiers on ChIP-seq data

Note that only the best hit per ChIP-sequence is considered in the current version of the module.


The module requires:

  • python2.7 (and does not work with python3 in its current version).
  • the BioPython module
  • the TFFM package accessed from your PYTHONPATH environment variable.
  • the scikit-learn module.
  • the pandas module.
  • access to bigWig files providing the values of the DNA shape features HelT, MGW, ProT, and Roll from your genome interest along with the second order computation these features. Please visit the GBshape website.
  • the bwtool.


You can find some examples of how to run the tool in the script provided in the test/ repository of this package.

The script plots the heatmap(s) of trained classifier(s). Note that the current version only works for PSSM/TFFM + DNA shape classifiers. You can get help on how to use it by typing

python2.7 -h

Project home page

For information on the source tree, examples, issues, and pull requests, see


If you use the DNAshapedTFBS tool, please cite

  • A. Mathelier, B. Xin, T.-P. Chiu, L. Yang, R.R. Rohs, and W.W. Wasserman (2016) DNA shape features improve transcription factor binding site predictions in vivo. Cell Systems, DOI:10.1016/j.cels.2016.07.001.