Judging a Structure Determination

The use of the Rfree has had a near revolutionary effect on protein structure refinement. It is now de rigueur to report Rfree values for any protein structure refinement. The only point that remains in contention is what value constitutes an appropriate Rfree, given a particular resolution, data quality, molecular weight, and so on. 

In evaluating a structure determination for accuracy and precision, it is usually prudent to inquire as to the quality of the electron density map according to which the final model was built, and to the properties and quality of the final model. The former question can be addressed in real space, that is, how good is the electron density map, and/or in reciprocal space, that is, how well determined were the phases, and how well measured the structure amplitudes that contributed to the map. The model, of course, can be judged by how well it predicts the diffraction intensities (the R factor), by how it explains chemical and biochemical questions, and how well it agrees with canonical stereochemical properties, such as bond lengths and angles. 

If the electron density map was originally phased using isomorphous replacement or anomalous dispersion methods, then a number of statistical criteria are usually available, 

Ultimately the question of electron density map quality is answered by whether we can trace a single polypeptide or polynucleotide chain through the density in a manner consistent with the known amino acid or nucleotide sequence. In doing so, we consider the agreement between amino acid side chains and the density assigned to them, whether selenium atoms in a map experimentally determined by single (SAD) or multiple anomalous dispersion 

FIGURE 10.12 Computer graphics representation of the electron density, calculated at 1.4 A resolution, of an a helix from a bacterial xylanase. 

---