Training on 2015.03.30
FR-X generator:
XG Control tells you the settings values (45 kV and 66 mA vs 20 kV and 10 mA when not used) . You can screen as soon as the parameters have reached the target values, but wait one more hour to collect, so that the beam is stable.
Pilatus 300K detector:
>thread
command on camserver gives the temperature and humidity level Green light at the back of the detector to check before use.
Cryostream should be at 100 K
Close 3 glass doors, push the green light on the left 1)
Run HKL_3000_P300KChi 2)
Collect/Connect to initialize
(also open an instrument server window in the background)
0.25° oscillation by default As a “rule of thumb” set the detector distance (in mm) to half the longest unit cell parameter (in Å) 3)
Align/Zero goniostat
resets all positions (parameters in /hkl_dc.ini)
2θ always swings away (clockwise)
Set the directory /data/user/project1_crystal1
Never collect at 2θ = 0, use at least 2θ = 2 to offset the center horizontally and avoid thus avoid missing data due to the gaps between detectors
You can swing the slider on Time window to adjust/reduce the number of collected frames Don't forget to adjust the Mosaicity value! If anomalous data is required, adjust the redundancy accordingly
For some pins, Chi 60° may be too high and should be restricted to 50°
Close the window and choose Setup DC
to transfer the information to HKL data collection window
Exposure time is not imported from the strategy and should adjusted manually
multi mode is useful to find the best detector distance and exposure time
Check diffraction images go to the data directory
dtdisplayhkl *.img &
30 to 50 strong reflections are needed for indexing. If it is not the case, increase the exposure time. It's possible to index non consecutive images if they have the same scan id
At 2θ ω goes from -90° to +90°, so don't put 90° step with 0° as ω start
Open display, do the peak search there, frame ↑ (middle click), the peak search has automatically changed, OK
Chi geometry goniometer, omega angle 180° follows the 2θ angle
If you press ABORT, the connexion will be lost and you will need to re-initialize Check the log on the MSCservdect window. To resume, check the parameters in the display. The scan number will be implemented.
With a 3D window >1, the mosaicity box will automatically be checked
Box size: (16 by default) usually not necessary to change, even if there are neighbor spots in the box Int_box: red ellipsoid to integrate spot. The reject area is between the red and the white. Is it necessary to increase the size for the red ellipsoid (0.35 by default)?
Changing box size also changes the spot size. Use the box size that gives the highest resolution for the red area (it depends if the number is odd or even).
Profile fitting radius is the yellow circle. When ON, it display only the strongest reflexions. Increase (from 8 to 10) when there are less spots to have enough strong reflexions within the circle, otherwise HKL might reject too many weak reflexions. Don't use a too big radius because spot profile is different at the center and in the edges.
3D window: number of consecutive images put together to estimate the mosaicity. Usually it covers the spread of the mosaicity ONCE, but set it to cover it TWICE.
Set shadow region: the Mask for the beamstop is not critical.
All scans are processed χ² should be close to 1
Check the boxes:
Have low reflexions marked for rejection. Scale several times until the number don't change anymore. If χ² is too low, decrease the error scale factor. Exclude frames if necessary.
For SAD phasing: if χ² is high: increase the scale factor (up to 1.6) and the error model (up to 0.6-0.7), but it might kill the anomalous signal. Check the orange(merged)/blue(unmerged) plot on χ² : higher signal means higher gap
Use auto correction: will try to reach a χ² of 1, but we don't know how, so be careful.
The log file contains merged I+/I- an the end of the file and tables with unmerged I+/I- in the middle of the file.
The sequence is necessary. Give the source or the signal or define if you have a heavy atom. Software used: Arp/Warp, CCP4-6.3.0, CCP4-6.4.0(Coot), ShelX
Data analysis: Grey area 1.3 for anomalous signal (can be used up to 1.0)
Find sites: Choose the resolution limit (>1.8 Å to distinguish disulfide bonds, test 0.1 Å increments) Check the absolute value and the spread of solutions (a few in the up-right corner)
Site view: Sphere size correlates with the occupancy
Phase: High resolution limit; try different resolution cutoffs
Check hand (ShelXE): You should see the difference, dark red/dark blue, difference of one graduation in scale. If necessary, increase the cycle number so that a plateau is reached.
Phase: Mlphase, DM-phase extension
NCS: DM or Parrot (a bit better) At least try it if you have more than 1 molecule in the Asymmetric Unit
Build: ARP/WARP for high resolution Buccaner for lower resolution (∼ 3 Å) Let ARP/WARP build/refine several times if has difficulties building. Start several times witha few cycles rather that once with many cycles.
Analyze data
Prepare Model: Edit the PDB file (delete ligands, waters, side-chains if you want)
Run MR (MolRep): 3 Å by default; rigid body at 3.0 Å, then refine at max resolution