Best B-factor for map sharpening

Hi there, very inexperienced user here and processing data for the first time. Many thanks for this great software and the very active community.

It is recommended to try out different B-factors (plus visual inspection) for sharpening maps after refinement (non-uniform in my case), as previsouly discussed somewhere else on this board (Safe map sharpening).

Would this be equivalent now to export the final volume (map file, not map_shap), for instance, to the “Autosharepn Map” tool from the Phenix package (where anisotropy correction is also performed) and let it make the best objective decision?

Thanks in advance.

Hi @ALBA,
Thanks for posting and sorry for the delayed response!

You definitely can take the xx_map.mrc file (i.e. not the sharpened map) and use it in downstream tools like Phenix, including for auto-sharpening. Note that the xx_map.mrc file from a refinement job is filtered using the global gold-standard FSC curve, to remove information at resolutions beyond trustable resolutions. Due to this filtering, post-hoc sharpening will generally be “safe” in the sense that it can only change how the refined structure information is visualized, but can’t sharpen information that should not have been trusted in the first place.

Note that the refinement jobs also output unsharpened, unfiltered half-maps from the gold-standard refinement. These can be used to also perform resolution estimation or other post-processing in other tools.