Fig. 10
From: Cobdock: an accurate and practical machine learning-based consensus blind docking method
The binding site identification and pose prediction performance of CoBDock for two proteins, 1T4E and 3MXF. A and C represent the cavity and the ligand pose on proteins, 1T4E and 3MXF. The proteins are coloured by magenta, and the structure in cyan represents the cavity found as the top prediction. B and D represent the natural ligand and prediction of ligand pose, in green and yellow respectively