The K-Pax page

This software contains an implementation of a Bayesian model-based method for simultaneously classifying aligned proteins into functionally divergent subgroups and identifying their function specific residues. Download the K-Pax software package .

The software code was written by Pekka Marttinen, and the Bayesian model has been developed jointly by Pekka Marttinen, Jukka Corander, Liisa Holm and Petri Toronen. Details of the method are given in the paper:

Marttinen, P., Corander,J., Törönen, P. and Holm, L. (2006). Bayesian search of functionally divergent protein subgroups and their function specific residues. Bioinformatics, 22, 2466-2474.

An example data set containing 296 sequences from the Urease protein family is included in the K-Pax package. Supplementary material about the identified function specific residues for this data set can be downloaded Download here .

The K-Pax software is currently designed for Windows 2000/XP and requires installation of Matlab runtime component (downloadable here). We aim to provides upgrades for Mac OS X and Linux systems during Spring 2007.