OK, it doesn't happen if I generate maps as part of a refinement run
using ccp4i - they overlay with the Coot-generated mtz maps perfectly.
So for the moment I'll assume I screwed up when I ran fft, and I'll now
go back and try and see what I did wrong. I'll also make sure that I
check things more carefully before asking stupid questions.
Eleanor Dodson wrote:
On Mar 4 2008, Craig Morton wrote: This is rather bad - I presume in
both cases you are reading the same coefficients? FWT PHWT for instance?
Eleanor
Morning,
in some cases we're finding substantial differences in what we see if
we allow Coot to generate density automatically from an mtz as opposed
to generating maps with CCP4 and reading them in - specifically some
large peaks in a CCP4 difference map at about 3.5 to 4 sigma that are
effectively non-existant in the Coot-generate mtz difference map
(partly visible at 2.2 sigma but really just part of the noise). This
is for a pretty well refined 1.7A dataset.
Is there a consensus amongst the Coot-community about mtzs vs. maps?
Do users regard Coot-generated mtz density as good for quick
exploration but use maps for more detailed work? Or am I being overly
trusting of CCP4?
Cheers,
Craig.
--
Dr Craig Morton Ph: +61 3 9288 2480
Senior Associate Fx: +61 3 9416 2676
St Vincent's Institute www.svi.edu.au
