Running the DPS/Mosflm GUI

Scaling and Merging

This system is for processing images from the ADSC Quantum-4 CCD detector. In theory, it should also work for ADSC Quantum-1 images, but this has not been tested. Modification to the command files, and perhaps to source code, would be needed to process images from other detectors - this is a project for the future.

To process a set of data, start up dpg_motif_gui. Use the "Get Cmd Files" button to copy command files from their standard location to your working directory. If you have previously processed data in the same directory, you can skip this step. Then you must load a parameter file. If you don't have one from a previous run, use "prototype.param". Make changes in the main window and popup windows (get them by using buttons at top right) as needed for your data and save in an appropriately named parameter file. Then follow the flow chart across the middle of the GUI: initialization, indexing, refinement, integration, scaling, final output.


DPS indexing works on the first image in the set. (If you want it to use some other image, temporarily change the "Start Image No." to the number of that image). If provided with the correct distance, wavelength, and direct beam position, it generally gives a good solution. To use it, follow the sequence of buttons across the top of the window which appears when you click "Index:DPS". "Find Peaks" starts up the adxv display program. Be sure the cross indicating direct beam position is on the direct beam spot. If not, move it using the "Settings" item of the "Edit" menu. Use the "Predictions" item of the "View" menu to bring up a window with a "Peak Search" button. Click on it to find peaks. The arrows to the right of the button let you request more or fewer peaks. After finding peaks, exit from adxv. Then click "Autoindex". Possible solutions in different symmetries will be listed. If you had not already put the correct symmetry in the "Cell Parameters" window, do so now. Click "Save Result". Then click "Display Result". Type "@display.inp" in the Xmosflm window. When the image appears, click "Predict" to see predicted spot positions. After checking predictions, click "EXIT", answer questions (no, you don't want to save information in a mosflm save file, and yes, you do want to exit), and type "end" in the Xmosflm window. If prediction is not satisfactory, check distance, wavelength, and direct beam position, and you may want to pick more or fewer peaks. Varying the "Maximum Cell Edge" value may help. Most predictions must overlap spots, or refinement will fail. Close the autoindexing window when done with it.

Mosflm indexing uses one or more images. It is less reliable than DPS indexing for a single frame but will generally succeed if you can give it enough images which are separated by at least several degrees in spindle angle. To use, click "Index:MOSFLM" and type "@index.inp" in the Xmosflm window. In the mosflm GUI which comes up, use "Find spots" to locate peaks. Answer questions in the xdl_io_window, accepting defaults. If there are too many or too few peaks picked, adjust the "Threshold" value and try again. Then click "Autoindex", and answer questions. If no good solution is found, use "Read image" to read another image in the series, use "Find spots" for the new image, and use "Autoindex" using spots from both images. If there is still no good solution, spots from more images can be included. When you get a reasonable result, accept it and use "Predict" to see predictions. You may need to adjust distance, wavelength, or direct beam position - set these in tha parameter list running down the left side of the window. Cell parameters from a mosflm indexing must be transferred to the "Cell Parameters" window of the main GUI manually, also any changes in distance, wavelength, or direct beam position that you may have made. Click "EXIT", answer questions, and type "end" in Xmosflm window when done with indexing.


Refinement of distance, direct beam position, and other "camera constants" is automatically carried out during integration. However, it is usually desirable to refine unit cell parameters and mosaicity in a preliminary run before integrating the whole set of images. There are two ways to do this:

1) Mosflm refinement using the "POSTREF SEGMENT" option. To do this, bring up the "Refinement Control" window and select images for "Mosflm Refine". Pick two sets of 2-4 images separated by at least 10 degrees or so in phi (if available). Select whether to refine "Beam Divergence", i.e. mosaicity plus real beam divergence. "ISOTROPIC (MOS)" is usually appropriate. Then click "Refine:MOSFLM" in main GUI window. Results will appear in the Refinement Control window (after some time - there is no feedback of progress during the refinement). If they look reasonable, click "Save Beam Div ONLY" to save divergence/mosaicity, "SAVE Cell ONLY" to save new unit cell. You can also save the refined missetting angles, but this is generally not advisable.

2) Postchk refinement. To do this, first integrate 10-20 frames with no post-refinement: in "Additional Parameters", disable post-refinement. Set the starting and ending image numbers to include 10-20 frames. Then click "Integrate:MOSFLM". Integration will be done in the background, with no notice until it finishes. In the Refinement Control window, select the frames for Postchk (the ones you just integrated), and click "Postchk" in the main window. If results in Refinement Control look good, save the cell parameters, and perhaps beam divergenece/mosaicity.


In "Additional Parameters" window, set any necessary parameters. Post- refinement should be enabled unless it proves to be unstable, but cell parameters should not normally be refined. Mosaicity should be refined only if you think it really does change during the run. You may need to change the spot separation criteria if your spots are closely spaced. Defaults for other parameters are probably fine. For an explanation of the meaning of the various parameters, bring up an Xmosflm window (with Index:MOSFLM), type "help", and follow through the various menu items. Set starting and ending image numbers to cover the range you want to integrate, and click "Integrate:MOSFLM". The integration may take some time. To follow its progress, you can look at the log file which is being generated (...mosflm.log). When completed, a summary will appear in the main GUI window.

Scaling and merging:

Click on "Scaling:SCALA" to scale and merge output from integration. Look in the ...scala.log file for statistics on the process. Finally the "I to F:TRUNCATE" button will produce final files of F's, in both binary and ASCII formats.


Print out and read "mosflm_user_guide.doc" for help on using mosflm. Also, whenever an Xmosflm window is up, you can access mosflm interactive help by typing "help". To follow what the GUI program is doing in detail, look at the command files,, etc. Various input and log files are produced when the program is run; examination of these may help.

Revised: Tuesday, 13-Jan-1998 15:47:06 EST