Skip to main content
. Author manuscript; available in PMC: 2005 Sep 13.
Published in final edited form as: Bioinformatics. 2004 Aug 4;20(Suppl 1):i77–i85. doi: 10.1093/bioinformatics/bth912

Fig. 5.

Fig. 5

ROC plot—sensitivity versus specificity—for the present algorithm applied to three subfamilies, and the average ROC plot (solid curve). Sensitivity is the fraction of catalytic residues captured with a given threshold (cutoff) score, and specificity is the fraction of catalytic residues (TPs) among all identified (conserved) residues (TPss and FPs). On the average, a sensitivity of 1 is accompanied by a specificity of 0.6, i.e. two additional conserved residues are identified with the three catalytic residues. However, the accuracy of the method is remarkable, given that the five conserved residues (comprising the catalytic triad) are detected amongst N ~ 250 residues.