When running a Non-uniform refinement job, the output includes the unsharpened map (_map.mrc) and the sharpened map (_map_sharp.mrc). I occasionally have users choose to model against the unsharpened map and would like to clarify a few points for our documentation:
- Is a mask applied to map.mrc and if so which one?
- Which FSC curve would be appropriate to describe the resolution in the unsharpened map.mrc?
- Which mask is used to sharpen the _map_sharp.mrc map?
Thanks!
Hi @s_cryoEM!
Thanks for your enquiry about masking and FSC curves.
In answer to your questions:
- The unsharpened map is not masked
- The resolution of the sharpened and unsharpened maps is identical. When it comes to deciding which FSC curve you want to use to report the map resolution it can help to think about and assess the masks that are used to calculate them. Typically people may use the Tight FSC curve that uses the mask_fsc_auto, but it is always worth checking if this curve shows signs of being overly tight (see this tutorial) and checking that the mask_fsc_auto in ChimeraX appropriately covers your density. If the mask appears too tight then you can make yourself a better mask using Volume Tools, then calculate the FSC using this new mask, using the Validation FSC job.
- The mask_fsc_auto is used to calculate the B-factor used for sharpening but the sharpened map is not masked.
If you find that the unsharpened map is better for model-building than than the sharpened map from refinement, you may also like to try a Sharpening Tools job with a custom B-factor that is smaller than the value used in the refinement sharpening.
I hope that helps!