Hi,
for example, these qualities below will tell me something about your model:
1. Rfree, Rwork overall and shown in resolution shells;
2. Geometry statistics:
- overall rmsd's bonds, angles, ...;
- Molprobity statistics/scores;
- histograms of deviations from ideal bonds, angels, ...;
3. Plot of local density correlation (map CC) reported per residue/atom,
as well as mFo-DFc and 2mFo-DFc map values reported per residue/atom;
4. POLYGON picture showing where your model stands w.r.t. similar ones
available in the database (Acta Cryst. D65, 297-300 (2009)).
For example, all the above you can get in PHENIX using "Comprehensive
validation" tool.
Pavel.
On 1/18/10 8:00 AM, james09 pruza wrote:
> Dear all,
>
> What are the standard methods for the crystallographic model judgment?
> What parameters should be included in the final refinement statistics
> in relation to the model error? What is Cruinkshank DPI and how can it
> be calculated and what information it gives about the quality of the
> model?
>
> Thanks in advance for the suggestions.
>
> J.
|