Rebuilding the model at atomic resolution

Apart from the rebuilding of the protein and the solvent model in the same way as it is done for structures at less than atomic resolution, the major part of manual modeling at atomic resolution concerns the introduction of multiple discrete conformations. Accurate modeling of multiple conformations, in particular in active sites, often leads to much cleaner electron density and increased interpretabiUty. 

In this context, it should be stressed, that it is normally more efficient to wait with building multiple conformations until after anisotropic displacement parameters have been refined, as the inclusion of these parameters often leads to a dramatic improvement of map quality. 

Before submitting an newly created ins-file for a complete run, it may be helpful to run one cycle of refinement against a fake reflection file containing only 100 reflections—such a run will not spend much time on analyzing the diffraction data but quickly generate the full geometrical analysis of the model which can be used to find out where mistakes have been made in the manual rebuilding. 

To identify regions of the structure that need rebuilding, the SHELXL Ist-file contains 

In most cases, atoms that have been marked as non-positive definite should be removed from the model and possibly be reintroduced after some cycles of refinement if difference density persists. 

---