Figure 1

Methods for defining interaction patterns. (A) As an example of fragmentation of amino acid residues, the case of alanine is illustrated, where eight fragments were obtained: N-CA-CB, C-CA-CB, C-CA-N, O–C-CA, and the fragments with the reverse orders. (B) Contacting pairs of a protein fragment and a ligand atom were collected from the dataset. When the distance between the first atom of the fragment and the ligand atom was less than the criterion (the sum of van der Waals radii plus the offset value 1.0 Å), the pair was sampled. Sampled contacting pairs formed a 3D spatial distribution on the basis of reference coordinates defined by the three atoms in the protein fragment. Three arrows indicate the axes of the reference coordinate system, “CA-C-O” is a protein fragment, and “C3” in circles denote positions of interacting ligand atoms observed in the dataset (“C3” means a type of ligand atom i.e., a sp3 carbon atom). (C) A pattern recognition technique was applied to each 3D spatial distribution. Interaction patterns were defined as Gaussian mixture distributions.