[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [ccp4bb]: averaging
*** For details on how to be removed from this list visit the ***
*** CCP4 home page http://www.ccp4.ac.uk ***
Dear Bernard and Gokulan,
Have you tried using DM or RESOLVE to improve your phases? Both can start
from a simple model (with DM use lsqkab to get NCS operators) and perform
iterative rounds of NCS averaging and solvent flattening, which helps if you
have a poor starting model. I have started to do this for regular model
building, after CNS refinement. I take a Fo.Fc.phs file from cns and convert
it into an ascii file with
Fo Fc sFc and PhiC
which I then convert into an MTZ file for resolve. Resolve also needs a few
atoms per monomer to be able to calculate the NCS operators. Resolve will the
automatically generate a mask based on your ED (not as optimal as a manually
built mask).
Here is an example script for running RESOLVE:
#!/bin/sh
# Here is a very minimal script to run prime-and-switch phasing
#
# Set CCP4 variables for symmetry information and
# for file handling:
#
#ccp4_setup
#setenv SYMOP /usr/local/lib/solve/symop.lib
#setenv CCP4_OPEN UNKNOWN
##############################################################################
# Use the phase file from CNS (Fo sFo Fc PhiC)
##############################################################################
.../pmb_cns2phs.pl fo.sfo.fc.fob
##############################################################################
# Convert PHSS file to MTZ
##############################################################################
f2mtz HKLIN new_s.phs HKLOUT new_s.mtz << f2mtz_eof |tee f2mtz.html
TITLE MTZ from PHSS Dps T2b3 From FoFc_calc.inp -> pmb_cns2phs.pl
PNAME D175D3_50
DNAME D175D3_50
CELL 80.109 80.109 132.296 90 90 120
LABOUT H K L FO SIGFO FC PHIC
CTYPout H H H F Q F P
FORMAT '(3F5.0,4F11.3)'
SYMM 154
f2mtz_eof
##############################################################################
# Use SIGMAA to determine phase probabilities
##############################################################################
sigmaa HKLIN new_s.mtz HKLOUT new_sa.mtz << simaa_eof |tee sigmaa_log.html
TITLE SIGMAA m*Fo-Fc map
###RESOLUTION 500.0 1.9 ! Resolution limits in Angstroms
RANGES 20 5000 ! Number of bins for analysis v. resolution
#### ! Monitor every 5000th reflection
PARTIAL ! Option for difference map coefficients
ERROR ! Use sigmaA v resolution for coordinate error
LABIN FP=FO SIGFP=SIGFO FC=FC PHIC=PHIC
END
simaa_eof
#
# create the HA.pdb file for NCS
#
grep " CA ARG [A-C] 28 " minimize.pdb > ha.pdb
grep " CA ALA [A-C] 138 " minimize.pdb >> ha.pdb
grep " SD MET [A-C] 185 " minimize.pdb >> ha.pdb
#
##############################################################################
# Now run prime-and-switch phasing:
##############################################################################
resolve <<EOD |tee resolve_map.log
hklin new_sa.mtz
labin FP=FO FC=FC PHIC=PHIC FOM=WCMB FWT=FWT
hklout resolve.mtz
solvent_content 0.47 ! your solvent content goes here (lower by 5%
to avoid trimming ED).
prime_and_switch
EOD
# Now "resolve.mtz" has the output amplitudes, phases,
# figure of merit and HL-coeffs in columns labelled: FP PHIM FOMM HLAM HLBM
HLCM HLDM
#
mtz2various HKLIN resolve.mtz HKLOUT resolve_map.phs << mtz2_eof
LABIN FP=FP FOM=FOMM PHIC=PHIM HLA=HLAM HLB=HLBM HLC=HLCM HLD=HLDM
OUTPUT USER '(3I5,3F10.3, 4F8.3)'
mtz2_eof
#
echo "Run xfft to create an FSFOUR map file from resolve_map.phs"
#
# Make a fin file from the model phas file to be phased by the resolve map in
xmergephs
#
awk '{printf"%5d%5d%5d%10.3f %8.3f%10.3f %8.3f \n",$1,$2,$3,$4,9.0,$5,9.0}'
083017_.06.pmb_lscale.phs > 083017_.06.pmb_lscale.fin
#
Bernhard Rupp wrote:
> *** For details on how to be removed from this list visit the ***
> *** CCP4 home page http://www.ccp4.ac.uk ***
>
> Dear All:
>
> Gokulan and I are stuck with what ought to be a simple procedure (harhar)
> Given: a h k l fo, fom, phwt file, warped from a trimer w/o using NCS
> Wanted: a 3-fold averaged map with NCS, in hope to improve weak density.
> Succeeded:
> make a ccp4 map (awk, fft)
> make a cell map (mapmask)
> make a xtalview map (map2fs, thanks Frank!)
> rtftm maprot (t for thin)
> used NCS ops from CNS (probably questionable as the rmsd is 4.5A)
> nevertheless, script made and resulting map has no overlap according
> to log.
> Inspection of input maps satisfactory.
> Transposition of matrix gives same nonsense. Maps look unchanged from input
> map.
>
> I guess the result is Bad news:
>
> Mapping density from cell using operator 1
> Mapping density from cell using operator 2
> Mapping density from cell using operator 3
> Mapping density to cell using operator 1
> Density correlation: 0.00000 over 0: 0 0: 0 0: 0
> Mapping density to cell using operator 2
> Density correlation: 0.00000 over 245:245 -113:*** 113:113
> Mapping density to cell using operator 3
> Density correlation: 0.00000 over -68:-68 -181:*** 175:175
>
> Obviously, we are doing something wrong.
> Secondly, is there a better way to get consistent NCS matrices from direct
> manipulation
> of the maps?
> and third, Should a mask be made instead of just copying the CNS ncs
> matrices?
>
> thx, Gokulan and br
>
> INPUT
>
> Data line--- CELL 83.123 104.872 105.196 90.000 90.000 90.000
> Data line--- GRID 240 240 240
> Data line--- MODE BOTH
> Data line--- SCALE 0.33333
> Data line--- AVER
> Data line--- ROTA POLAR 0.0 0.0 0.0
> Data line--- TRANS 0.0 0.0 0.0
> Data line--- AVER
> Data line--- ROTA MATRIX -0.44014 0.63215 -0.63770 -0.62202 0.29754
> 0.72426 0.64759 0.71544 0.26225
> Data line--- TRANS 84.76557 -49.18177 49.70381
> Data line--- AVER
> Data line--- ROTA MATRIX -0.51502 -0.63090 0.58027 0.58774 0.23286
> 0.77482 -0.62395 0.74009 0.25088
> Data line--- TRANS -23.52714 -79.15917 76.67988
>
> is assembled from CNS NCS ops (get_ncs_martix.list)
>
> ======================================================================
>
> NCS operator which transforms:
> ( segid B ) -> ( segid A )
>
> matrix= ( -0.44014 0.63215 -0.63770 )
> ( -0.62202 0.29754 0.72426 )
> ( 0.64759 0.71544 0.26225 )
> translation= ( 84.76557 -49.18177 49.70381 )
>
> rms difference= 4.29163
>
> ======================================================================
>
> NCS operator which transforms:
> ( segid B ) -> ( segid C )
>
> matrix= ( -0.51502 -0.63090 0.58027 )
> ( 0.58774 0.23286 0.77482 )
> ( -0.62395 0.74009 0.25088 )
> translation= ( -23.52714 -79.15917 76.67988 )
>
> rms difference= 4.10269
>
> ======================================================================
>
> -----------------------------------------------------------------
> Bernhard Rupp
> Macromolecular Crystallography and Structural Genomics
> LLNL-BBRP L448 Phone (925) 423-3273
> University of California Phax (925) 424-3130
> Livermore, CA 94551 email br@llnl.gov
> URL http://www-structure.llnl.gov
> TB Structural Genomics Consortium http://www.doe-mbi.ucla.edu/TB
> EU Mirror http://www.ccp14.ac.uk/ccp/web-mirrors/llnlrupp
> -----------------------------------------------------------------
--
Best Regards,
Mark
-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_
Mark A. White, Ph.D. Tel:
Sealy Center for Structural Biology (409) 747-4747
Dept. of Human Biological Chemistry & Genetics Fax:
Basic Science Building, Room 660C (409) 747-4745
University of Texas Medical Branch email:
Galveston, TX 77555-0647 mailto://white@bloch.utmb.edu
http://xray.utmb.edu http://www.xray.utmb.edu
-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_