An evaluation of selected density-modification methods for protein-structure determination

The effectiveness of density-modification methods in reducing phase errors has been quantitatively analyzed under conditions in which the source and magnitude of phase error could be defined. Progress was gauged by monitoring both the phase errors and electron density. Changes in electron density were monitored in and around the protein density and in the solvent density. The results of this analysis suggest that the improvement possible is more dependent upon the magnitude than on the source of phase error, and that the methods are most effective when the initial phase error is relatively small. As the magnitude of error increases, the improvement gained decreases, and in some cases noise is introduced by the procedure(s).