Learn how to visualize the reflections of your diffraction dataset.
Reflections with I/SigI > 2 are deemed to be good data
for structure solution. This is often used as a resolution cutoff limit. Here
we examine a dataset to see that there isn’t always a well defined resolution
The dataset “merged.mtz” provided by a workshop participant Armando Albert is opened in the HKLviewer.
The IMEAN,SIGIMEAN reflections are shown by double clicking the row with that label in the upper left table.
To better see any variation in the IMEAN data values either untick the “User defined power scaling” checkbox on the Sizing tab or click on the colour chart to open the Colour Gradient Maps dialog. Then select a smaller power factor value than 1, say 0.29.
Right click the first row with the “IMEAN,SIGIMEAN” label
and select “Make new data as a function of this data and another dataset”.
In the “Create New Reflection data” dialog enter “newdata=flex.abs(data1)/sigmas1” and “IoverSIGI” as the formula and the label for the new data respectively.
In the text field python syntax can be entered for the predefined variables:
|python syntax||predefined variables|
|newdata||miller_array._data variable produced by the operation|
|newsigmas||miller_array._sigmas variable produced by the operation|
|data1||miller_array._data variable of the reflection right-clicked for bringing up the dialog used by the operation, labelled “FP” in Figure 2|
|sigmas1||miller_array._sigmas variable of the reflection right-clicked for bringing up the dialog used by the operation, labelled “SIGFP” in Figure 2|
|data2||miller_array._data variable of the dataset selected from the drop-down list, labelled “FWT,PHWT” in Figure 2|
|sigmas2||miller_array._sigmas variable of the dataset selected from the drop-down list if it exists.|
The input syntax is plain python executed as one command using the flex array API defined by the cctbx. Multiple commands can be separated with semi-colon. Basic knowledge of flex array data types and the relevant API is assumed. When the operation succeeds the newly created array will be displayed in the viewer.
Execute the formula and display the IoverSIGI data set in the viewer by clicking OK:
On the Binning tab, bin data against IoverSIGI.
Change the number of bins from 1 to 5 or 6.
HKLviewer will try to put an equal number of reflections into
each bin. But here we want explicit bin thresholds.
Enter bin thresholds by typing in numbers 4,3,2,1 from the bottom row and up in the “lower bin value” column.
Untick the checkboxes of the first two bins to hide reflections with I/SigI less than 2.
This proves to be a significant number of reflections not just located beyond a certain resolution limit. The relatively small number of reflections with I/SigI > 2 suggests that it could be difficult to solve the structure for this data set.
tNCS is a condition where two or more molecules in the unit cell are aligned in a way to cause interference in the diffraction pattern. It shows as periodic modulation of the diffraction pattern and can cause problems for structure solution unless identified and taken into account by structure solution software. tNCS vectors are real space vectors and can be predicted by in advance of structure solution. The dataset 6C5F (available from the PDB) has tNCS which is detected by programs such as phenix.xtriage. It determined the tNCS vector to be 0.333, -0.334, 0.040 in fractional coordinate space.
Open the data file in the HKLviewer.
Double click the FP,SIGFP label to display it.
On the Expansion tab, expand the reflections to a sphere.
On the sizing tab, tick the “User defined power scaling” checkbox and enter a power factor of 2 to show reflections roughly proportional to original intensities (which are not present in the file from the PDB).
On the Vectors tab, enter the coordinates for the tNCS vector found by xtriage on the last line by overwriting the “new vector” label to “xtriageTNCS” and entering the actual coordinate in the same row under the “as abc” column. Pressing return will submit the vector to HKLviewer.
Tick the xtriageTNCS vector checkbox as well as the “Orient vector” checkbox and the radio button “Parallel to screen”.
Either rotate the reflections manually on the slider control or tick the “Rotate continuously” checkbox.
This will make visible the modulations in the data as ripples in the reflections in planes perpendicular to the tNCS vector. It can be useful to tick the “Show Real Space Unit cell” checkbox to see where exactly the tNCS vector points.
Phasertng, which is currently under development, stores a data column, TEPS, revealing the modulation of the intensities. The coordinate for the tNCS vector is also embedded in the mtz file. This is illustrated below where the reflections have been sliced with a clip plane width of 8. On the sizing tab we have set the power factor to 0 and the linear scaling factor to 0.8. The modulation is clear from the alternating layers of red, green and blue layers of reflections.
The phenix tutorial with the file, porin.mtz, illustrates twinning. The twin operator found by phenix.xtriage is -h-k,k,-l.
Load the file in the HKLviewer and display the dataset “F-obs,SIGF-obs”.
On the Vectors tab, Enter the twin operator into the table of vectors; first by overwriting the last row with the “new vector” string with “twin”. Then by entering the string “-h-k,k,-l” in the adjacent rotation column and pressing return for the HKLviewer to accept the new twin operator. It turns out to be a 2-fold rotation so the HKLviewer will prefix the name with 2-fold.
Press the “Orient vector” checkbox and the radio button “Normal to screen”.
On the Expansion tab, expand the reflections to P1 and Friedel mates.
On the Slicing tab, tick the “Slice with a clip plane” checkbox and set “Distance from Origin” and “Clip plane width” to 4 and 1 respectively.
Show a “Tabulated Reflection data” window by right clicking the column label “F-obs,SIGF-obs” in the upper left table and selecting “Show a table…” from the menu.
To inspect data values of reflections related by the twin operator right-click a reflection in the view. It will bring up two tooltips with the hkl indices at the end of two vectors from the reciprocal space origin pointing to the twin reflections. Due to the clip plane, the vectors may not be fully visible. Double click one of the reflections to bring up the standard tooltip with all data values represented by this reflection. Right-clicking a reflection will also highlight the data value in the “Tabulated Reflection data” window. The HKL indices in the table are the original from file and so may differ from computed indices of expanded reflections in the tooltips displayed in the view.