ANNOUNCE: Jnet secondary structure prediction method

[James Cuff]

                 /* Announcing Jnet v0.1 (alpha) */

Jnet is a new consensus neural network protein secondary structure
prediction method.  The software is freely available to all.

The method works through the application of multiple sequence alignments,
alongside PSIBLAST and HMM profiles.  Consensus techniques are applied to
predict the final secondary structure more accurately than the component
predictions.
  
- Average Q3 prediction accuracy is 76.4% (8.4sd).  This figure is based
  on a set of 406 non-redundant proteins that were not applied during
  the development of the method.

- Residues predicted with a confidence value of 5 or greater have an
  average Q3 accuracy of 84%, and cover more than 68% of residues.
 
- Relative solvent accessibility based on a two state model, for 25, 5
  and 0% accessibility is predicted at 76.2, 79.8 and 86.6%
  respectively.

- Availability:

  The Jnet homepage can be accessed at:

  http://circinus.ebi.ac.uk:8081/jnet

  and from the software pages of the Barton group web server at the EBI:

  http://barton.ebi.ac.uk/

  SGI and Linux binaries are available, along with the C source code.

  Full details of the method are described in: "Application of enhanced 
  multiple sequence alignment profiles to improve protein secondary
  structure prediction", Cuff, J. A., Barton, G. J., (submitted) 1999.


James Cuff,
Geoff. Barton
 
European Bioinformatics Institute, 
Wellcome Trust Genome Campus,
Hinxton, Cambridge. CB10 1SD