advertisement: compare things at compare-stuff.com!

Structure-mapping of sequence information

Whilst carbon-$\alpha$ coordinates usefully define the backbone of a fold, many analyses of protein structure[Hendlich et al., 1990] and threading algorithms[Jones et al., 1992] use carbon-$\beta$ coordinates. In helices and strands the position of the carbon-$\beta$ contains more information about the alternating side-chain locations and environments than the carbon-$\alpha$ coordinates. It was an oversight not to use carbon-$\beta$ coordinates in the mapping of sequence and structure, although for structure alone carbon-$\alpha$ coordinates are a reasonable choice because they represent the backbone trace through space (the `fold').

Two interesting applications for self-organising maps were suggested in Section 5 involving the identification of active sites and insertion points for protein engineering. Hopefully, these will be pursued at a later date. The representation of folds in two dimensions could also be developed, the first step would be to display the secondary structures on the map trajectories.