A simple method is described for identifying well-defined regions in a set of protein structures calculated from experimental interproton distance restraints. Two different functions, one based on the mean global rms difference, the other on the distance variation between equivalent atoms in different residues, are used to distinguish ‘variable’ from ‘well-defined’ regions. These functions are calculated in an iterative manner. The method is also capable of identifying several locally well-defined regions whose relative positions are not well-defined globally.
