I am wondering how the "goodness" scale for density fit analysis is determined. I have a case here where to the eye there is a very good fit of model to density throughout 99% of the unit cell, yet the validation tool marks most residues with a big red bar.
Is there some hidden property [e.g. map gradient?] that triggers a bad validation score even though the fit to density looks fine by visual inspection at a single contour level? Is the good-to-bad scale normalized to the individual structure, so that a 'good everywhere' structure looks 'uniformly bad' in the validation tool? Would the presence of large difference peaks outside the area of the current model make any difference? -- Ethan A Merritt Biomolecular Structure Center University of Washington, Seattle 98195-7742
