SigDis - Signature discovery in biosequences


This page is under construction!


SigDis is a supervised learning algorithm that finds patterns (signatures), in the form of regular expressions, that discriminate two sets of unaligned protein or DNA sequences.

SigDis searches for patterns in two phases: first some anchors (words overrepresented in the sequences) are quickly discovered. Then, we search for a pattern(s) using the an- chors found as seeds of a pattern search, thus focusing the search of the patterns in potentially interesting regions of the sequences.


We are currently finishing the first release of SigDis. It will be available here soon for Linux platforms. For the moment it is available solely upon request to Nuno Fonseca.


References to SigDis

  • Nuno A. Fonseca, Cristina P. Vieira, Peter W. Holland, and Jorge Vieira. Protein evolution of ANTP and PRD homeobox genes. BMC Evolutionary Biology, 2008.
  • Jorge Vieira, Nuno A. Fonseca, and Cristina P. Vieira. S-RNase based gametophytic self-incompatibility system evolved only once in eudicots. Journal of Molecular Evolution, 2008.