Hi Elena, Cross correlation coefficients are frequently used to describe the quality of fit of an atomic model in an EM map. Other measures such as the number of atoms at positions with density above 1 sigma are also used. The Chimera fit in map tool in recent Chimera versions (snapshot or daily build since Feb 20, 2008) can report correlation coefficient. Correlation measures the agreement between a simulated map at a specified resolution with the experimental map. You enter that resolution in the fit map dialog in the options panel. Also turn on "use map simulated from atoms". Then press the "Update" button on the line of the dialog that says "correlation" and "average map value". The fitting also reports the number of atoms lying outside the displayed map contour surface. You can set that to 1 sigma or 2 sigma or a value of your choosing. To figure out the standard deviation of the map values use volume dialog menu Tools / Volume Mean, SD, RMS. More fit map documentation is in the manual http://www.cgl.ucsf.edu/chimera/docs/ContributedSoftware/fitmaps/fitmaps.htm... The volume guide also has info but is not up to date with the latest Chimera features: http://www.cgl.ucsf.edu/chimera/tutorials/volumetour/volumetour.html#fitmap Tom

Hi Hernando, If the fit in map option to "Use only data above contour level from first map" is enabled (it is by default) then the correlation is calculated using only grid points within the displayed contour level of the first map. If you turn that option off, then the correlation uses all grid points of the first map. If the second map contains other structures (e.g. other asymmetric units) within the bounds of the first map box, then using all grid points will give a lower correlation due to the mismatch between the maps in those regions. So it is usually best to use the default setting and set the contour level of the first map low enough to include all grid points where the correlation is desired. Tom hsosa@aecom.yu.edu wrote:

One thing I am curious about it, when calculating correlation values, is how to deal with cases in which the pdb atomic structure of only one asymmetric unit is fitted into an electron density corresponding to more than one asymmetric unit. Is there an easy way to exclude from the calculation the densities corresponding to the extra asymmetric units.

Thanks

Hernando

Thomas Goddard wrote:

...

Hi Elena,

Cross correlation coefficients are frequently used to describe the quality of fit of an atomic model in an EM map. Other measures such as the number of atoms at positions with density above 1 sigma are also used.

The Chimera fit in map tool in recent Chimera versions (snapshot or daily build since Feb 20, 2008) can report correlation coefficient. Correlation measures the agreement between a simulated map at a specified resolution with the experimental map. You enter that resolution in the fit map dialog in the options panel. Also turn on "use map simulated from atoms". Then press the "Update" button on the line of the dialog that says "correlation" and "average map value".

The fitting also reports the number of atoms lying outside the displayed map contour surface. You can set that to 1 sigma or 2 sigma or a value of your choosing. To figure out the standard deviation of the map values use volume dialog menu Tools / Volume Mean, SD, RMS.

More fit map documentation is in the manual

http://www.cgl.ucsf.edu/chimera/docs/ContributedSoftware/fitmaps/fitmaps.htm...

The volume guide also has info but is not up to date with the latest Chimera features:

http://www.cgl.ucsf.edu/chimera/tutorials/volumetour/volumetour.html#fitmap

Tom _______________________________________________ Chimera-users mailing list Chimera-users@cgl.ucsf.edu http://www.cgl.ucsf.edu/mailman/listinfo/chimera-users