BCL::Contact is a novel contact prediction method that utilizes artificial neural networks (ANNs) and specializes in the prediction of medium to long range contacts (contacts between amino acids far apart in the sequence). BCL::Contact only utilizes sequence information and has individiual ANNs specialized for helix-helix, helix-strand, strand-helix, strand-strand and sheet-sheet contacts. A consensus prediction is calculated by merging the outputs from these 5 ANNs.
The output file has an extension .contact but can be viewed with any text editor. This file contains for every residue pair x and y in the sequence; residue id for x, one letter code for x, residue id for y, one letter code for y, contact predictions from helix-helix, helix-strand, strand-helix, strand-strand, sheet-sheet, and the merged consensus prediction.
Additional options available in the full BCL release. Licenses (including the binary) for academic (free of charge) or commercial users (a fee applies) can be obtained from http://bclcommons.vuinnovations.com.