Hello,
I am experiencing strange behavior in coot when trying to create anomalous
difference maps from ccp4-produced mtz files. It appears to be a bug in coot
but it is possible that it is caused by a ccp4 program. It seems to be the same
problem as was reported a year ago
(http://www.mail-archive.com/coot@jiscmail.ac.uk/msg01122.html) but does not
appear to have been resolved. In this case I have 2 mtz files that have been
created through identical paths (as far as I can tell!) and one works and one
doesn't in the same version of coot and on the same machine. Hopefully this
information will help in identifying the source of the problem.
I have an mtz file containing dano and phwt columns (created by using cad to
merge the phases from refmac with the original mtz file from truncate that
contains dano). From coot, I open the mtz, select expert mode, select dano for
the amplitudes and select fwt for the phases and then OK. Usually this will
open up a perfectly good anomalous map with nice peaks on seleniums, etc. but
now I have an empty map with no density - the contour level reads: nan3/A^3 (
nansigma). The terminal window looks like this:
(filter-fileselection-filenames-state)
(manage-column-selector /ccp4/dano-t.mtz)
Making map from /1646b2/1646b2/DANO_1646b2inf and /1646b2/1646b2/PHWT
making difference map from MTZ filename /ccp4/dano-t.mtz using
/1646b2/1646b2/DANO_1646b2inf /1646b2/1646b2/PHWT
CCP4MTZfile: Mtz column type mismatch: DANO_1646b2inf D-F
Number of OBSERVED reflections: 46057
INFO:: finding ASU unique map points with sampling rate 2
INFO grid sampling...Nuvw = ( 324, 324, 324)
INFO:: 0.6 seconds to read MTZ file
INFO:: 0.009 seconds to initialize map
INFO:: 1.474 seconds for FFT
INFO:: 0.65 seconds for statistics
Map mean: ........ nan
Map sigma: ....... nan
Map maximum: ..... -1e+10
Map minimum: ..... 1e+10
INFO:: 0.026 seconds for contour map
INFO:: 2.759 seconds in total
(make-and-draw-map-with-reso-with-refmac-params "/ccp4/dano-t.mtz"
"/1646b2/1646b2/DANO_1646b2inf" "/1646b2/1646b2/PHWT" "" 0 1 0
"Fobs:None-specified" "SigF:None-specified" "RFree:None-specified" 0 1 0 -1.00
-1.00)
several things about this strike me as a potential problem;
1) CCP4MTZfile: Mtz column type mismatch: DANO_1646b2inf D-F
2) Map mean: ........ nan
Map sigma: ....... nan
3) " 0 1 0 "Fobs:None-specified" "SigF:None-specified"
"RFree:None-specified"
I try another mtz file created exactly the same way but with a different
dataset to compare the terminal output with an mtz file that worked earlier for
this same project and it looks like this:
(filter-fileselection-filenames-state)
(manage-column-selector /ccp4/job17-peak-dano.mtz)
Making map from /1646b2/1646b2_2/DANO_1646b2pk and /1646b2renamed/1646b2/PHWT
making difference map from MTZ filename /ccp4/job17-peak-dano.mtz using
/1646b2/1646b2_2/DANO_1646b2pk /1646b2renamed/1646b2/PHWT
CCP4MTZfile: Mtz column type mismatch: DANO_1646b2pk D-F
Number of OBSERVED reflections: 41811
INFO:: finding ASU unique map points with sampling rate 2
INFO grid sampling...Nuvw = ( 324, 324, 324)
INFO:: 0.445 seconds to read MTZ file
INFO:: 0.009 seconds to initialize map
INFO:: 0.451 seconds for FFT
INFO:: 0.08 seconds for statistics
Map mean: ........ -5.51326e-08
Map sigma: ....... 0.0100791
Map maximum: ..... 0.0944656
Map minimum: ..... -0.0629124
INFO:: 0.004 seconds for contour map
INFO:: 0.989 seconds in total
(make-and-draw-map-with-reso-with-refmac-params
"/sgpp/data/process/Lmaj00847AAA/SeMet/1646b2/ccp4/job17-peak-dano.mtz"
"/1646b2/1646b2_2/DANO_1646b2pk" "/1646b2renamed/1646b2/PHWT" "" 0 1 0
"Fobs:None-specified" "SigF:None-specified" "RFree:None-specified" 0 1 0 -1.00
-1.00)
this time there are values for the map mean and sigma and I get a beautiful
anomalous difference map. Again there are what I thought were signs of a
problem (the column type mismatch and the None-specified column labels) in the
terminal output so points 1 and 3 above are not the culprit. This suggests that
there is a problem with the latest mtz file but it appears to be essentially
the same as the earlier one that works so...???
Further confusing the matter, when I use fft in ccp4 to create an anomalous map
using the "problematic" mtz file and the same DANO_1646b2inf and PHWT columns
(which was the suggested "fix" to the problem last year!), it yields an
anomalous map that opens up just fine in coot with the expected peaks on the
seleniums suggesting that the "problematic" mtz file is fine - at least to ccp4
that is.
The "work around" works fine but it is kind of annoying to have to do the extra
fft step. Hopefully this information is useful.
Please let me know if you need more information.
Eric
__________________________
Eric Larson, PhD
Biomolecular Structure Center
Department of Biochemistry
Box 357742
University of Washington
Seattle, WA 98195
