[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[ccp4bb]: AMoRe - Help needed with Translation function using E's in non-orthogonalspace groups
*** For details on how to be removed from this list visit the ***
*** CCP4 home page http://www.dl.ac.uk/CCP/CCP4/main.html ***
[Posted on behalf of David Borhani - please reply to
<borhand@basf-corp.com>
(not to me! thanks...)]
*************************************************
Dear CCP4 Bulletin Board Members,
I cannot get the AMORE translation and fit functions to work with E's, rather
than F's, when running in space groups P1 and P2(1), using command procedures
that have worked well for years on many different crystals (all orthorhombic or
tetragonal). No trouble in either case with the cross rotation function...works
great.
My command procedure was originally based on an example in the CCP4
distribution...basically calculate F's from the search model (SFALL), convert to
E's (ECALC), then create the packed HKL and table files using AMORE, with the
SORTFUN and SORTFUN MODEL keywords, respectively. Then use these files in the
AMORE CRF and TF. When I do this in orth. or tetrag. crystals, have always
gotten the right answer very easily. In P2(1) (several cases), the TF has never
worked. I have in the past resorted to TFFC, using E's, which has always worked
great.
In a recent P1 crystal (two molecules in asymmetric unit, related by ~2-fold
axis...crystal is nearly monoclinic), I saw the same behaviour: AMORE TF did not
work at all, but TFFC did (both run with E's). BUT, when I ran AMORE again with
F's, using a AMORE table generated straight from the search model (i.e. F's),
the TF worked, giving the same answer as TFFC.
Finally, I re-ran AMORE with E's, using a modification of my old procedure
described above wherein I let AMORE move the search model to where it thought
best (TABFUN NOTAB), used these re-oriented search model coordinates to
calculate F's, E's, packed HKL's, and a new table. Re-ran the CRF, worked fine.
Re-ran the TF, using the new CRF results (different from before, because AMORE
rotated my search model), and it did NOT work.
I cannot figure out what I am doing wrong here, nor why this procedure works
fine with orth. or tetrag. crystals. Help would be appreciated. I have included
two command files below, first the run with F's (worked), and then the run with
E's (did not work).
Thanks very much.
David Borhani
Group Leader, Biochemistry Dept.
BASF Bioresearch Corp.
100 Research Drive
Worcester, MA 01605 USA
borhand@basf.com
508-849-2944
---------------------- Amore TF with F's (works) -----------------------
!/bin/csh -f
# tabling run for model
amore xyzin1 search_model_cent.pdb \
table1 search_model_cent.tab \
> amore_trans2.log \
<< EOF_tab
TITLE tabling for search_model_cent.tab
VERBOSE
TABFUN NOROTATE NOTRANSLATE
MODEL 1 BREPLACE 0 BADD 0
SAMPLE 1 DMIN 3 SHANN 2 SCALE 4.0
END
EOF_tab
#
# sorting run for Fobs of unknown crystal
amore HKLIN ../../xray_data/crystal.mtz \
HKLPCK0 crystal.hkl \
>> amore_trans2.log \
<< EOF_pack
TITLE packing h k l
SORTFUN RESOL 100. 3.
LABIN FP=F SIGFP=SIGF
EOF_pack
#
# Now calc. translation function.
amore hklpck0 crystal.hkl \
table1 search_model_cent.tab \
mapout amore_trans2.tf \
>> amore_trans2.log \
<< EOF_tra
TRAFUN CB NMOL 2 RESO 8 3.25 PKLIM 0.5 NPIC 10
SYMM 1
VERBOSE
TITLE TF molecule 1 fixed at origin; where is molecule 2?
CRYSTAL ORTH 1 FLIM 0.E0 1.E8 SHARP 0.
SOLUTIONRC FIX 1 269.92 90.32 273.03 0.0000 0.0000 0.0000 31.6 1
SOLUTIONRC 1 90.18 89.89 272.78 0.0000 0.0000 0.0000 29.2 2
EOF_tra
#
exit
Above command procedure gives this output:
Best Solution: 0.760 0.401 0.500 16.460
scale, corr-amplitudes, r-factor, corr-intensities sol_nb.
0.14005E+01 60.897 36.995 60.059 1
Spacegroup P1: Correlations all equal -
Only Phased translation peak height significant
Translation solutions sorted by correlation coefficients
TABLE Alpha Beta Gamma Tx Ty Tz Corr_F Rfac Corr_I
Pkcount Dmin
SOLUTIONTF1 1 269.92 90.32 273.03 0.0000 0.0000 0.0000 31.6 1.0 0.0
1 0.0
Spacegroup P1: Correlations all equal -
Only Phased translation peak height significant
Translation solutions sorted by correlation coefficients
TABLE Alpha Beta Gamma Tx Ty Tz Corr_F Rfac Corr_I
Pkcount Dmin
+++++>>>> SOLUTIONTF2 1 90.18 89.89 272.78 0.7599 0.4008 0.4999 60.9
37.0 60.1 1 0.0 <<<<+++++
Normal Termination of TRAFUN
Some weird messages, I guess having to do with P1, but the correct answer, with
a Corr. Coeff on F of 60.9.
---------------------- Amore TF with E's (does NOT work) -----------------------
#!/bin/csh -f
# let amore move the model as it wants.
amore xyzin1 search_model.pdb \
xyzout1 search_model_reorient.pdb \
table1 /dev/null \
>> amore_trans3.log \
<< EOF_tab
TITLE model shift for search_model_reorient.pdb
! Don't actually make the table...will do below after E's are calculated.
VERBOSE
TABFUN NOTAB
MODEL 1 BREPLACE 0 BADD 0
SAMPLE 1 DMIN 3 SHANN 2 SCALE 4.0
END
EOF_tab
#
# calculate SF's for shifted model.
# amore has chosen a rather large cell nearly 200x200x200 angstroms
# this cell is quite similar to the one it chose in the F's procedure above.
#
sfall xyzin search_model_reorient.pdb \
hklout temp1.mtz \
>> amore_trans3.log \
<< EOF_sfall
MODE SFCALC XYZIN
SFSGROUP 1
SYMMETRY 1
RESOLUTION 1000. 2.7
TITLE SFcalc for search_model_reorient.pdb
END
EOF_sfall
#
# calculate E's...
ecalc hklin temp1.mtz \
hklout search_model_reorient.ecalc \
>> amore_trans3.log \
<< EOF_ecalc
TITLE Model (search_model_reorient.pdb)
REFLECTIONS 100000
RESOLUTION 0
LABIN FP=FC
SPACEGROUP P1
EOF_ecalc
#
# Produce HKLPCK0 and TABLE1 files from the E's...
amore HKLIN search_model_reorient.ecalc \
HKLPCK0 search_model_reorient.hkle \
SORTING_NR 2000000 \
>> amore_trans3.log \
<< EOF_pck
TITLE ** Packing h k l E for Model search_model_reorient.pdb **
SORTING
LABIN FP=E
EOF_pck
#
amore hklin search_model_reorient.ecalc \
table1 search_model_reorient.tabe \
TABLING_MI 5000000 \
TABLING_MR 5000000 \
>> amore_trans3.log \
<< EOF_tab
TITLE ** Tabling h k l E for Model search_model_reorient.pdb **
SORTFUN MODEL 30 2.7
LABIN FC=E PHIC=PHIC
EOF_tab
#
{... Here the CRF was calculated using E's...worked well, as it has previously
many times...}
{ E's for the Fobs were calculated previously, before running the SRF...}
#
# Now the TF with E's...
amore hklpck0 crystal.hkle \
table1 search_model_reorient.tabe \
mapout amore_trans3.tf \
TABLING_MI 5000000 \
TABLING_MR 5000000 \
>> amore_trans3.log \
<< EOF_tra
TRAFUN CB NMOL 2 RESO 8 3.25 PKLIM 0.5 NPIC 10
SYMM 1
VERBOSE
TITLE TF molecule 1 fixed at origin, where is molecule 2?
CRYSTAL ORTH 1 FLIM 0.E0 1.E8 SHARP 0.
SOLUTIONRC FIX 1 199.58 169.50 64.98 0.0000 0.0000 0.0000 0.2 99.2
-0.7 35.1 1
SOLUTIONRC 1 22.19 169.08 67.51 0.0000 0.0000 0.0000 0.3 99.2
-0.1 32.0 2
EOF_tra
#
exit
Above command procedure gives this (NOT WORKING) output:
Peaks must be greater than 0.71
The top 10 are selected for output
X,Y,Z translation, height, order
0.7353 0.8056 0.0000 0.14140E+01 1
0.7353 0.7222 0.0000 0.14140E+01 2
0.7353 0.6389 0.0000 0.14140E+01 3
0.7353 0.5556 0.0000 0.14140E+01 4
0.7353 0.4722 0.0000 0.14140E+01 5
0.7353 0.3889 0.0000 0.14140E+01 6
0.7353 0.3056 0.0000 0.14140E+01 7
0.7353 0.2222 0.0000 0.14140E+01 8
0.7353 0.1389 0.0000 0.14140E+01 9
0.7353 0.0556 0.0000 0.14140E+01 10
Best Solution: 0.735 0.806 0.000 1.414
scale, corr-amplitudes, r-factor, corr-intensities sol_nb.
0.67702E+00 0.673 98.492 -0.221 1
0.67702E+00 0.673 98.492 -0.221 2
0.67702E+00 0.673 98.492 -0.221 3
0.67702E+00 0.673 98.492 -0.221 4
0.67702E+00 0.673 98.492 -0.221 5
0.67702E+00 0.673 98.492 -0.221 6
0.67702E+00 0.673 98.492 -0.221 7
0.67702E+00 0.673 98.492 -0.221 8
0.67702E+00 0.673 98.492 -0.221 9
0.67702E+00 0.673 98.492 -0.221 10
Spacegroup P1: Correlations all equal -
Only Phased translation peak height significant
Translation solutions sorted by correlation coefficients
TABLE Alpha Beta Gamma Tx Ty Tz Corr_F Rfac Corr_I
Pkcount Dmin
SOLUTIONTF1 1 199.58 169.50 64.98 0.0000 0.0000 0.0000 0.2 99.2 0.0
3 0.0
Spacegroup P1: Correlations all equal -
Only Phased translation peak height significant
Translation solutions sorted by correlation coefficients
TABLE Alpha Beta Gamma Tx Ty Tz Corr_F Rfac Corr_I
Pkcount Dmin
+++++>>>> SOLUTIONTF2 1 22.19 169.08 67.51 0.7353 0.6389 0.0000 0.7
98.5 -0.2 3 0.0 <<<<+++++
Normal Termination of TRAFUN
Also, in contrast to the F's run, here the TF map has a peak of only 1.4, sigma
1.0 (F's, peak ~16, sigma 1).