The paper presents a novel method for verifying protein structures by analyzing the nonbonded atomic interactions between different types of atoms (carbon, nitrogen, and oxygen). The method classifies these interactions into six types (CC, CN, CO, NN, NO, and OO) and uses a quadratic error function to characterize the distribution of these interactions in a database of 96 reliable protein structures. By comparing the interaction patterns in candidate protein structures with those in the database, regions of incorrect or misregistered residues can be identified. The method is effective in distinguishing between correctly and incorrectly determined regions, with a sensitivity to errors in backbone positions on the order of 1.5 Å. The authors also apply the method to a survey of the Protein Data Bank (PDB), identifying structures that may contain errors. The FORTRAN program ERRAT is provided for further analysis.The paper presents a novel method for verifying protein structures by analyzing the nonbonded atomic interactions between different types of atoms (carbon, nitrogen, and oxygen). The method classifies these interactions into six types (CC, CN, CO, NN, NO, and OO) and uses a quadratic error function to characterize the distribution of these interactions in a database of 96 reliable protein structures. By comparing the interaction patterns in candidate protein structures with those in the database, regions of incorrect or misregistered residues can be identified. The method is effective in distinguishing between correctly and incorrectly determined regions, with a sensitivity to errors in backbone positions on the order of 1.5 Å. The authors also apply the method to a survey of the Protein Data Bank (PDB), identifying structures that may contain errors. The FORTRAN program ERRAT is provided for further analysis.