Figure 4
From: Functional site prediction selects correct protein models
"Summary plots" for five training-set proteins. The "summary plots" for each of the five training-set proteins (a) 3chy, (b) 1coz, (c) 2trx, (d) 1f4p and (e) 1di0 are shown. In each plot, the "Fold Score" is plotted against the measure of structural correspondence to the native protein. (Note, both these measures are plotted on reversed scales). The best models lie towards the lower left corner in each plot having a high score and high structural similarity to the native. The native structure itself is plotted as a large red dot and all folds that correspond to the native are also red with others blue. The different symbols designate different decoy fold families.