The following people and programs are very gratefully acknowledged:

• Sean Eddy: For the HMMER2 package [Although Sean did giggle when I said I was using HMMER for secondary structure prediction :-)]
• The Stuttgart Neural Network Simulator project: The Jnet networks were trained with this simulator - it is an excellent piece of software.
• Richard Copley - some of the conversion scripts started off life as hacked versions of Richard's Perl scripts.
• Steve Searle - helped enormously with aspects of the C code.
• David Jones - The PSIPred method influenced many of the ideas for this project. His enlightening discussions have helped enormously.
• Burkhard Rost - For PHD, and for giving me the faith that neural network prediction methods can and do work.
• I'd also like to thank the MRC and EMBL for their funding.