Hi Rakesh,

  Chimera Fit in Map and also the measure correlation command by default only compute the cross-correlation within the displayed envelope of the first map.  So it should give the masked value you are looking for, although the actual value will depend sensitively on the surface level of the first map that you choose since that will control which grid points near the boundary of your fragment are included.  There is no right answer about how tightly to mask and since the correlation value is so sensitive to that it is not too meaningful to try to compare correlation values although people do it anyways.

  Check out the measure correlation and fit in map docs for more details

    https://www.cgl.ucsf.edu/chimera/docs/UsersGuide/midas/measure.html#correlation

    https://www.cgl.ucsf.edu/chimera/docs/ContributedSoftware/fitmaps/fitmaps.html

  Tom


On 12/6/2021 5:09 PM, Rakesh Ramachandran wrote:
Sorry, what I meant was say I am fitting a fragment into a large map, the cross correlation should be calculated only around the fitted region of the fragment, that is any voxels outside the fragment location should not be taken into account in the cross-correlation score. 

I need this because I am fitting fragments of different sizes into a map and I tend to get quite similar cross correlation scores even though principally it should.

On Mon, Dec 6, 2021 at 5:03 PM Tom Goddard <goddard@sonic.net> wrote:
Hi Rakesh,

  I don't understand what you mean by "length normalized by masking".  Chimera can compute cross-correlation using the Fit in Map panel.  If you are fitting an atomic model to the map it needs to simulate a map from the atomic model  since correlation is between two maps.  You do that using the "Use map simulated from atoms" option in the Fit in Map panel Options pane.

    Tom


On 12/6/2021 4:00 PM, Rakesh Ramachandran wrote:
Hi Tom,
  
   I am using Chimera Fit in Map for EM fitting. Though it does fit well, I want to calculate a cross correlation score based on the model map size, that is the cross correlation should be length normalized by masking the rest of the map.

--
Regards
Rakesh Ramachandran



--
Regards
Rakesh Ramachandran