Figure 1.

Prediction of active sites. Panel A: The percentage of correctly predicted active sites in human monomeric and homodimeric proteins: blue bars refer to predictions relying on the identification of the most conserved pocket (DrosteP protocol), red bars refer to predictions relying on the identification of the largest pocket, green bars refer to predictions relying on the identification of the largest among the conserved pockets, violet bars refer to predictions relying on the identification of the most conserved pocket without DrosteP optimization, orange bars refer to predictions based on Jensen Shannon Divergence score. Panel B: Solvent accessible volume in protein pockets: total, conserved or non conserved ones.

Cammisa et al. BMC Bioinformatics 2013 14(Suppl 7):S9   doi:10.1186/1471-2105-14-S7-S9