Fitmap on overlap/correlation only within bounds of smallest map
Hi, I am trying to fit a set of large (whole protein) maps to a map corresponding to a single domain of the protein (extracted using ‘vop zone minimalbounds true’ with an atomic model of the domain in question. It works fine, except that it takes rather a long time (~2000 steps), and the final correlation is 0.33, which is strange because the overall correlation between the two complete maps is ~0.95, and I can see that the final overlap between the aligned maps is good. It seems like ‘fitmap’ is calculating the correlation/overlap based on all grid points in the map that I am aligning, which has a much larger box size than the map to which I am aligning it. Is there any way to restrict the metric analysis to just consider points within the smaller box of the reference map? I cannot find such an option in the documentation for fitmap. The use case for this is when one has multiple conformations of a protein (for example, EM volumes after 3D-classsification) and would like to align them on a single reference domain. Best, Oliver.
Hi Oliver, When fitting map A in map B Chimera by default uses the grid points inside the contour surface shown for map A. You can also choose to have it use all points in the box for map A. But there is no option to make it use the points inside the contour surface of map B other than fitting B in A — but that moves map B. I agree it would be nice to have such an option. I’ll make a feature request for it. So here is the tricky way to handle this now. Fit a copy of the single domain map into each of the full protein maps. then move the single domain map copies and their aligned full maps so that the single domain maps all align. For that last step you use the Chimera “matrixcopy” command. Suppose you fit domain map #0 into protein map #1 and another copy of the domain map #2 into protein map #3. Then use matrixcopy #0 #2 moving #3 to move domain copy #2 to overlap domain map #0 and carry protein map #3 along with the same motion. http://www.cgl.ucsf.edu/chimera/current/docs/UsersGuide/midas/matrixcopy.htm... <http://www.cgl.ucsf.edu/chimera/current/docs/UsersGuide/midas/matrixcopy.html> Tom
On Apr 9, 2015, at 8:13 AM, Oliver Clarke wrote:
Hi,
I am trying to fit a set of large (whole protein) maps to a map corresponding to a single domain of the protein (extracted using ‘vop zone minimalbounds true’ with an atomic model of the domain in question.
It works fine, except that it takes rather a long time (~2000 steps), and the final correlation is 0.33, which is strange because the overall correlation between the two complete maps is ~0.95, and I can see that the final overlap between the aligned maps is good.
It seems like ‘fitmap’ is calculating the correlation/overlap based on all grid points in the map that I am aligning, which has a much larger box size than the map to which I am aligning it. Is there any way to restrict the metric analysis to just consider points within the smaller box of the reference map? I cannot find such an option in the documentation for fitmap.
The use case for this is when one has multiple conformations of a protein (for example, EM volumes after 3D-classsification) and would like to align them on a single reference domain.
Best, Oliver. _______________________________________________ Chimera-users mailing list Chimera-users@cgl.ucsf.edu http://plato.cgl.ucsf.edu/mailman/listinfo/chimera-users
Thanks Tom, that makes sense, I will give it a try. Oli.
On Apr 13, 2015, at 5:18 PM, Tom Goddard <goddard@sonic.net> wrote:
Hi Oliver,
When fitting map A in map B Chimera by default uses the grid points inside the contour surface shown for map A. You can also choose to have it use all points in the box for map A. But there is no option to make it use the points inside the contour surface of map B other than fitting B in A — but that moves map B. I agree it would be nice to have such an option. I’ll make a feature request for it.
So here is the tricky way to handle this now. Fit a copy of the single domain map into each of the full protein maps. then move the single domain map copies and their aligned full maps so that the single domain maps all align. For that last step you use the Chimera “matrixcopy” command. Suppose you fit domain map #0 into protein map #1 and another copy of the domain map #2 into protein map #3. Then use
matrixcopy #0 #2 moving #3
to move domain copy #2 to overlap domain map #0 and carry protein map #3 along with the same motion.
http://www.cgl.ucsf.edu/chimera/current/docs/UsersGuide/midas/matrixcopy.htm... <http://www.cgl.ucsf.edu/chimera/current/docs/UsersGuide/midas/matrixcopy.html>
Tom
On Apr 9, 2015, at 8:13 AM, Oliver Clarke wrote:
Hi,
I am trying to fit a set of large (whole protein) maps to a map corresponding to a single domain of the protein (extracted using ‘vop zone minimalbounds true’ with an atomic model of the domain in question.
It works fine, except that it takes rather a long time (~2000 steps), and the final correlation is 0.33, which is strange because the overall correlation between the two complete maps is ~0.95, and I can see that the final overlap between the aligned maps is good.
It seems like ‘fitmap’ is calculating the correlation/overlap based on all grid points in the map that I am aligning, which has a much larger box size than the map to which I am aligning it. Is there any way to restrict the metric analysis to just consider points within the smaller box of the reference map? I cannot find such an option in the documentation for fitmap.
The use case for this is when one has multiple conformations of a protein (for example, EM volumes after 3D-classsification) and would like to align them on a single reference domain.
Best, Oliver. _______________________________________________ Chimera-users mailing list Chimera-users@cgl.ucsf.edu <mailto:Chimera-users@cgl.ucsf.edu> http://plato.cgl.ucsf.edu/mailman/listinfo/chimera-users
participants (2)
-
Oliver Clarke -
Tom Goddard