[3dem] Resolution estimate: FSC vs PDB modeling

[Penczek, Pawel A]

Hi,

the point I wanted to make may or may not be relevant to the current current discussion, but as
it is sometimes confused I would like to clarify it again.

The FSC estimate (comparison of two NOISY half-maps or two noisy maps in general)
is not numerically equivalent to what I call cross-resolution (FCR), which is a comparison of
a noisy map with a NOISE-FREE map.  This would be for example a comparison of an EM map with
a PDB-derived model, the latter for the sake of simplicity considered noise-free.

It can be shown that

FCR = sqrt(FSC)


In simple words, if one estimated EM resolution from half-maps using FSC and say 0.5 cut-off,
the cross-resolution FCR of the same map with a X-ray derived model should be 0.71 at the same
frequency to be equivalent.  If FCR is 0.5, the cross-resolution would point at much poorer agreement.

Again, if one uses 0.5 as a cut-off for FSC, one should use 0.71 for FCR; if 0.147265 for FSC, 0.383 for FCR,
and so on…

For FSC there is also factor of two hanging there.


For more details please see
Penczek, P.A.: Resolution measures in molecular electron microscopy. Methods Enzymol 482, 73-100, 2010.


Regards,
-
Pawel Penczek
pawel.a.penczek at uth.tmc.edu<mailto:pawel.a.penczek at uth.tmc.edu>