Page Comparison

CX data processing on MX

All data on the MX beamlines is processed with XDS. For more information on XDS you can read the XDS paper or check out the XDS wiki.

Accessing your data on the beamline

The best way to access and analyse your data on the beamline computers is via the Linux terminal. If you are new to Linux, check out the Beginners Guide to Linux {insert link to Linux beginners guide}

The auto processed data can be found in the following directory:

/data/$EPN/home/$PI/auto/dataset/

where $EPN is the experiment number and $PI is the principle investigators surname.

You should see a directory with the form /data/$EPN/home/$PI/auto/dataset/yourexperiment_1_###########

Be aware that a directory is generated every time you screen as crystal, as well as when you collect a dataset. So when processing data, make sure you are processing a full dataset, and not crystal screening. You can check this by checking the run number for your dataset in the processing page, which should be the final number in the folder name.

{Insert picture of processing page and data folder}

Analysing your data on the beamline

In your dataset folder you will find:

• The XDS processed hkl file for your data processed in P1

• with a multi-scan absorption correction

• : XDS_ASCII.HKL_p1

• The

• XDS processed hkl file for your data processed in P1 without an absorption correction: XDS_ASCII.HKL_p1_noscale

• Three folders of data which has had sadabs absorption correction applied: sadabs_w, sadabs_m, sadabs_s. Each will contain:

  • A log file with infomration about the absorption correction: sad.abs

  • A hkl file: sad.hkl

  • An autoprocess.cif metacif: eg. autoprocess_sadabs_w.cif_mx1

• The autoprocess.cif_mx1 or autoprocess.cif_mx2 metacif file which has all the collection specific information you will need for youf final cif.

• A log file with information to compare data quality for the different absorption corrections: rint_log.txt

• • This will contain the resolution, number of spots (#data), theoretical number of spots (#theory), completeness, redundancy, mean intensity, mean intensity/sigma (Mean I/s), R(int) and Rsigma.

  • This is a good way to assess which absorption correction gave has the best data.

  • In the below example we would choose the XDS_ASCII.HKL_p1 to use to solve the structure, because it has the lowest R(int) and the highest Mean I/s. The noscale dataset can definitely be disregarded because it has a very high R(int).

Image Added