Experimental Phasing

Experimental phasing at SFX is challenging but now becoming more and more accessible. Here you will learn how to extract anomalous signals and how to solve the structure de novo.

Merge without Friedel's law

We will merge our lysozyme dataset without Friedel's law, that is, in 422 not 4/mmm pointgroup.

Modify our merging script and save it as process_anom_scale.sh,

#!/bin/bash

inp=$1
out=${inp%.stream}-anom-scale.hkl # added "-anom"

pg="422" # was 4/mmm

source ~sacla_sfx_app/setup.sh

process_hkl -y $pg --scale --odd-only -i $inp -o ${out}1  &
process_hkl -y $pg --scale --even-only -i $inp -o ${out}2 &
process_hkl -y $pg --scale -i $inp -o $out &
wait

and run it.

sh process_anom_scale.sh 266711-266721-dirax.stream

Also modify our statistics script and save it as stat_anom.sh,

#!/bin/bash

inp=$1
inp1=${1}1
inp2=${1}2
basename=${1%.hkl}
fom="R1I R2 Rsplit CC CCstar CCano" # added CCano
pdb="../lysozyme-opt.cell"
pg="422" # was 4/mmm
highres="1.9"

source ~sacla_sfx_app/setup.sh

if [ ! -d "stat" ]; then
	mkdir stat
fi

for mode in $fom
do
    compare_hkl $inp1 $inp2 -y $pg -p $pdb --fom=$mode --highres=$highres --nshells=20 --shell-file="stat/${basename}-$mode".dat 2>>stat/${basename}.log
done
check_hkl -p $pdb --nshells=20 --highres=$highres -y $pg  --shell-file="stat/${basename}-shells".dat $inp 2>>stat/${basename}.log

and run it.

sh stat_anom.sh 266711-266721-dirax-anom-scale.hkl

CCano indicates some anomalous signals, although very weak.

$ cat stat/266711-266721-dirax-anom-scale-CCano.dat 
  1/d centre    CCano       nref      d / A   Min 1/nm    Max 1/nm
     1.166  0.1341869        338       8.58      0.387       1.944
     2.195  0.2082471        381       4.56      1.944       2.446
     2.622  0.1037050        399       3.81      2.446       2.799
     2.939  0.1599720        399       3.40      2.799       3.080
     3.198  0.1493716        395       3.13      3.080       3.317
     3.421  0.0919954        410       2.92      3.317       3.524
     3.617  0.1278523        403       2.76      3.524       3.710
     3.794  0.1092259        413       2.64      3.710       3.879
     3.956  0.1620875        409       2.53      3.879       4.034

Calculate anomalous difference map

Here we use ANODE, which is distributed with the SHELX suite, to calculate anomalous difference map.

First, convert the merged intensities into the sca format through the mtz format.

create-mtz 266711-266721-dirax-anom-scale.hkl
phenix.reflection_file_converter 266711-266721-dirax-anom-scale.mtz --sca 266711-266721-dirax-anom-scale.sca

You can use mtz2sca utility in the CCP4 suite instead of phenix.reflection_file_converter.

Download the referernce model PDBID 4YOP and save it as anode.ent. This model is used to calculate the phase.

Next, make an input file shelxc.inp for SHELXC. Since we are not going to run SHELXD, only the first three lines are necessary. The following lines are included just for completeness. For details, see SHELXC keywords and SHELXD manual.

CELL 1.77 78.92 78.92 37.88 90 90 90
SPAG P43212
SAD 266711-266721-dirax-anom-scale.sca
SFAC S
FIND 6
DSUL 4
MIND -3.5
SHEL 100 2.2
NTRY 10000

Run SHELXC followed by ANODE.

shelxc anode < shelxc.inp | tee shelxc.log
anode anode # add -d2.2 to set the high resolution limit at 2.2 A.

Locations and heights of peaks are printed and saved to anode.lsa.

 Averaged anomalous densities (sigma)

   7.54  SD_MET
   7.42  CL_CL
   5.99  SG_CYS

... (omitted) ...

 Strongest unique anomalous peaks

          X        Y        Z   Height(sig)  SOF     Nearest atom

 S1    0.34688 -0.06475  0.48307    8.61    1.000    0.113  SD_A:MET105
 S2    0.37861 -0.00792  0.63836    8.46    1.000    0.162  SG_A:CYS115
 S3    0.40410 -0.10549  0.31108    7.93    1.000    0.306  CL_A:CL1002
 S4    0.20502 -0.01705  0.24550    7.27    1.000    0.094  SG_A:CYS76
 S5    0.25505 -0.07289  0.62559    6.68    1.000    0.163  SD_A:MET12
 S6    0.16500  0.09116  0.30155    6.50    1.000    0.246  SG_A:CYS80
 S7    0.26595 -0.10897  0.91974    6.48    1.000    0.166  SG_A:CYS6
 S8    0.37493 -0.01954  0.69397    5.86    1.000    0.320  SG_A:CYS30
 S9    0.18133  0.07180  0.30027    5.68    1.000    0.167  SG_A:CYS64
 S10   0.28271 -0.12114  0.87932    5.33    1.000    0.276  SG_A:CYS127

The map is saved as anode.pha. You can open it in Coot. You have to open the model file anode.ent first so that Coot knows the unit cell parameters.

Solving a structure de novo by Cu-SAD

We use a Cu-SAD dataset at CXIDB #34.

TODO Write!