SYSTEMS AND METHODS FOR MAKING ASSIGNMENTS IN ISOTOPE-LABELLED PROTEINS USING NUCLEAR MAGNETIC RESONANCE DATA

摘要

Computing systems and methods for characterizing a protein are provided. Each residue in a subset of the protein is in an amino acid type set and is represented by a vertex in a graph G formed from an atomic model of the protein. NMR data, acquired with some of the residues of the protein isotopically labeled, is used to form a graph H with each vertex representing a different residue of the protein and assigned one or more amino types. Placements of H onto G are formed, each including mappings assigning vertices in H to vertices in G subject to the constraints that vertices in H mapped to vertices in G cannot be of different amino acid types and edges between pairs of vertices in H must map to corresponding edges in G. For each vertex in H, the number of different valid mappings to G is determined by polling the placements as a constraint satisfaction problem and is deemed assigned when only a single unique assignment is identified.