CX Data Merging
- MX Cluster (Unlicensed)
Dataset merging - automated tools
How to use merging from the processing web interface.
On the Processing webpage select the datasets you want to merge. One must be selected as the reference dataset.
Click the blue Merge Selected Datasets button. A new line will pop up on the processing page, this is the merged dataset. Wait for the status to show Success before moving to the next step.
Access your data, which will be in /data/home/$EPN/$PI/auto/merger/$reference_dataset_name.
$EPN is your experiment number
$PI is the name of the user
$reference_dataset_name is the name of the dataset you selected as the reference on the processing page.
Within the data folder, there will be a folder called xscale, which contains the merged hkl, called xscale.hkl Open this hkl in xprep. When you are asked the hkl file format, select X for XDS_ASCII.HKL format from XDS or XSCALE.
|
|---|
Xprep option for reading in merged hkl |
Dataset merging - Manually
Merging with XSCALE
You can manually merge with XSCALE. Merging datasets that have been autoprocessed is a simple matter of copying the output XDS_ASCII.HKL_p1 files from their directories, renaming them as you go, then merging.
Each of the multiple collections will be auto processed as normal. Using the terminal find your way to the auto directory cd /data/home/$EPN/$PI/auto/dataset/
ls
You should see the directories
yourexperiment_1_###########
yourexperiment_2_###########
yourexperiment_n_###########
Make a new directory and copy with renaming the XDS_ASCII.HKL_p1 files
mkdir yourexperiment_merge_xscale
cp yourexperiment_1_###########/XDS_ASCII.HKL_p1 yourexperiment_merge_xscale/run1.hkl
cp yourexperiment_2_###########/XDS_ASCII.HKL_p1 yourexperiment_merge_xscale/run2.hkl
cp yourexperiment_3_###########/XDS_ASCII.HKL_p1 yourexperiment_merge_xscale/run3.hkl
Change to your new directory and merge the files
cd yourexperiment_merge_xscale
xscale2.py run1.hkl run2.hkl run.hkl
This will give you the merged and scaled data in P1. The program used is XSCALE and it is part of the XDS package with a log file called XSCALE.LP. You can now process XSCALE.HKL with xprep to determine the lattice, spacegroup and write an ins file and merged hkl file to solve your structure. (TIP, when using xprep, remember that XSCALE.HKL is in the XDS format [X], not the default HKLF4 format [4] ).
Merging With sadabs
You can use xds2sad on your XDS_ASCII.HKL_p1 file to generate a xds.sad file. The problem with this is that there is a multi-scan correction applied by XDS on the CORRECT data.
To merge reflection files that have not had the any scaling applied you need to use the xds.sad file. Make a new directory and copy with renaming the sadabs readable files with renaming.
mkdir yourexperiment_merge_sadabs
cp yourexperiment_1_###########/xds.sad yourexperiment_merge_sadabs/run1.sad
cp yourexperiment_2_###########/xds.sad yourexperiment_merge_sadabs/run2.sad
cp yourexperiment_3_###########/xds.sad yourexperiment_merge_sadabs/run3.sad
Change to your new directory and merge the files
Merging with Xprep
Each of the multiple collections will be auto processed as normal. Using the terminal find your way to the auto directory:
cd /data/home/$EPN/$PI/auto/dataset/
Find the directories of your data:
yourexperiment_1_###########
yourexperiment_2_###########
yourexperiment_n_###########
Run Xprep for each of your datasets, naming the output files run1, run2 and run3. Make sure they all have the same cell and space group. Note the unit cell of run1 (open the prp file and leave it open until later).
Make a new directory and copy run1.hkl, run2.hkl and run3.hkl files into it:
Change into the new directory and open xprep
start xprep with run1.hkl, you will need to enter the cell manually.
Go through xprep as normal, until you get to “Read, modify or merge datasets”. Instead of selecting [S], select [R] to read in another dataset. Select [O] to read in original unit cell, then enter then name of the hkl of the second dataset, run2.hkl. Select [R] again to read in the third dataset, the same as the with run2.hkl.
Once all three are read in, select [L] for “least square scale and merge datasets”. Select Linear or isotropic scaling then enter each dataset when prompted. When you have entered all the datasets, hit enter.
VERY IMPORTANT: when you see this line, you must select [U]. If you don’t, your data will have no Rint.
Continue working through xprep as normal. This will produce a merged hkl and ins file, use these to solve the the structure.