Figure 7
From: Protein-protein docking using region-based 3D Zernike descriptors
Given a receptor R and a ligand L in their original orientation, a random orientation T 0 is first applied to the ligand. Selection of a three-point basis (two points + the average of the normals of the two points) for both the ligand (L" → TIL) and receptor (TIR) allows coordinates to be expressed in the new reference frame. Alignment of the matched set of coordinates in the hash space then produces an additional transformation (L"' → T2L). To move into the original coordinate system, the transformation matrix Z is applied to the ligand protein.