Next: Data processing
Up: Web-Ice
Previous: Image Viewer
Subsections


Autoindex and strategy calculation

The Autoindex Tab is used to autoindex test diffraction images (with the option to collect the test images too) and to calculate data collection strategies that optimize data completeness. It can also be used to inspect a data collection strategy generated during sample screening.

The autoindex and strategy results for each sample are grouped under as a Run. Each run has got a unique name. The autoindex Tab navigation toolbar contains links to display a summary of all the user's runs (under user Runs), start an interactive application to generate new runs (New Run) or inspect or edit one of the runs (Selected Run)

Figure 4: Navigation toolbar in the Autoindex tab
Image autoindex-nav

The Runs page

The Runs page displays a table summarizing all the autoindexing results generated by Web-Ice.

Figure 5: Autoindex tab page showing a summary of all autoindex runs.
Image autoindex-allruns

For each run, the following are listed:

  1. The images used for autoindexing.

  2. The crystal score calculated as $1.0 - (0.7 \times e^{-4/d}) - (1.5 \times rmsr) - (0.2 \times \mu)$, where $d$ is the resolution limit in Å, rmsr (described above) is in mm, and $\mu$ is the mosaicity in degrees.

  3. The number of spots in the image and the number of reliable diffraction spots (found after rejecting the weakest and strongest spots).

  4. The number of ice rings.

  5. The autoindex software best guess of the crystal lattice.

  6. The estimated diffraction resolution limit based on $I/\sigma I$ statistics.

The runs can be sorted by alphabetical order or by time of creation. The Update reloads the page (this is useful to display results generated by Web-Ice in the background since the last visit to the page)

Selecting a run

Clicking on a run name allows the user to edit and reuse the run or inspect the results. Once a run has been selected, it will be highlighted in the user Runs table. The user can revisit a selected run by clicking on the name, or using the Selected Run link in the navigation toolbar.

Deleting runs

To clear the run and all its contents, click the link [Delete] in the Commands column of the user Runs table.


Generating a new run

Selecting New Run link in the Autoindex navigation toolbar opens a new run form page. The user has three options: 1) autoindex existing images stored in the MC computers, 2) import a strategy calculated during screening of the currently mounted sample, or 3) to collect and autoindex new test images from a mounted sample. The last two options require the user to connect to a beamline and they can only be used during beamtime.

A name for the run must be chosen. The run name must be a single word of alphanumeric characters.

Figure 6: New run form page.
Image autoindex-runform

Pressing the enter key or clicking the Create button will create the run; clicking on the Cancel button will return to the user Runs page.

The Run generation tool requires javascript enabled on the browser.


Autoindexing existing images

This option assumes that the user has already used Blu-Ice to collect the diffraction images and, for anomalous dispersion experiments, the fluorescence scans required to characterize the sample. In this case Web-Ice can read these files and supply a data collection strategy to the user.

The option is selected by clicking the ''Autoindex existing images'' button in the new run form page and clicking the Create button. This will take the user to the Setup page which guides the user through the following run set up steps (The Prev and Next buttons can be used to navigate between the various steps):

  1. Choosing the images: The directory and image names can be entered by the user or selected using the Browse button.

    Figure 7: Selection of directory and images for autoindex.
    Image autoindex-setupfiles

  2. Choosing the strategy: The user has the option to select the strategy using beamline specific parameters; this option is disabled unless the user has selected a beamline; this option takes into consideration the detector position and energy limits enabled at the beamline, and uses the beam intensity to provide an estimate of the absorbed dose. The option not to calculate the strategy is useful if the user does not intend to collect data from the crystal, but only examine the diffraction quality (e.g., during manual sample screening).

  3. Choosing the type of experiment: The types available are Native, MAD (three wavelengths) and SAD (SAD is also the recommended mode to collect data for Isomorphous Replacement phasing). Note that two oscillation ranges maximizing the unique and anomalous completeness will be always calculated regardless of the experiment type. The range optimizing anomalous completeness will be selected by default for MAD and SAD experiments.

    Figure 8: Selection of MAD experiments type for ''Autoindex Only'' option.
    Image autoindex-setupexp

  4. For MAD and SAD experiments, the energies may be input by the user, or imported from an autochooch summary file; use the directory browser tool to locate the summary file, select it by clicking on the radio button next to the file name and then click on the Load Scan File button. Important: The File filter in the Preferences tab must contain the entry *summary for the directory browsing tool to display this file type.

  5. To do a two-wavelength MAD experiment, change the appropriate energy input box to 0 (it is best to leave out the ''peak'' or the ''inflection'' energy - the remote energy should always be collected on MAD experiments).

    It is also possible to enter the file name (specifying the full path name in the input text box next to the Load Scan File button. Note that only the ''peak'' energy needs to be entered for a SAD experiment.

  6. The element and edge input boxes will be read in from the summary file. Note: If a summary file is not given, it is recommended to supply the heavy atom type. This makes it possible to give a more accurate estimate of the radiation dose for the experiment.

  7. Other options: The user can select the Laue group and cell if known. If given, the program will calculate the strategy only for that space group. If not given, the strategy will be calculated for the lowest space group belonging to the predicted Bravais lattice.

    For SAD and MAD experiments it is possible to enter the number of residues and heavy atoms in the monomer. If given, and the atom type is known, they will be used in the dose calculation for the strategy and also the exposure time and total number of images (based on the estimated signal).

    The page also displays the program used to calculate the starting and total rotation angle. BEST is used to calculate the full oscillation range for data collection; for multicrystal data collection, MOSFLM is used instead.

    Figure 9: Symmetry and strategy options.
    Image autoindex-setupother


Collecting test images and autoindexing

This option allow the user to collect test images and a fluorescence scan from a sample. In the current version, the sample must be previously mounted by hand, or using the robot via Blu-Ice, and centered by the user. The slit size should also be set to an appropriate value before hand.

The option to collect images and autoindex can be selected from the New Run form after choosing an accessible beamline from the toolbar. Creating the run will take the user to the Setup pages. The Prev and Next buttons can be used to navigate between the following steps:

  1. Choose sample: In the current version of Web-Ice, the user may only collect images from samples which have been already mounted either by the robot (via Blu-Ice) or manually. In later releases, it will be possible to use the robot to mount samples from Web-Ice.

  2. Choose Directory and image root name: The directory where the test images will be written to can be typed in or, if it already exists, selected using the Browse tool. The image root name will default to the run name.

  3. Choosing the strategy: The user has the option to select the strategy using beamline specific parameters; this option takes into consideration the detector position and energy limits enabled at the beamline, and uses the beam intensity to provide an estimate of the absorbed dose. The option not to calculate the strategy is useful if the user does not intend to collect data from the crystal, but only examine the diffraction quality (e.g., during manual sample screening).

  4. Choose the type of experiment. For SAD and MAD experiments, there is the option to collect a fluorescence scan from the anomalous element of interest; to do this, select the Perform Fluorescence scan option, and click on the absorption edge of interest in the periodic table appearing at the bottom of the page.

    For SAD experiments, the periodic table will enable selection of elements without an absorption edge in the beamline spectral range. Clicking on the element name will select the lowest energy attainable at the beamline for maximizing the anomalous signal present in the data. Important: This choice of energy does not guarantee that the anomalous signal can be detected. In some cases, collecting high redundancy data will also be necessary; in the current version this can only be done by manually extending the phi range supplied by Web-Ice.

    Figure 10: Selection of absorption edge to do a fluorescence scan
    Image autoindex-setupexp2

    If the option to collect the scan is not used, the energies may be input by the user or imported from an autochooch summary file; use the directory browser tool to locate the summary file, select it by clicking on the radio button next to the file name and then click on the Load Scan File button. You can also enter the file name (specifying the full path name in the input text box next to the Load Scan File button. Note: only the ''peak'' energy is required for a SAD experiment.

    To do a two-wavelength MAD experiment, change one of the energy input boxes to 0 (it is best to leave out the ''peak'' or the ''inflection'' energy - the remote energy should always be collected on MAD experiments).

    Note: If a fluorescence scan is not collected and summary file is not given, it is recommended to supply the heavy atom type. This makes it possible to give a more accurate estimate of the radiation dose for the experiment.

  5. Choose other options: The user can choose the exposure time, oscillation angle, beam attenuation and resolution for the test images. The default exposure time is a reasonable value for well diffracting crystals at the selected beamline. The default resolution is based on the current detector sample and energy (or the absorption edge energy for SAD or MAD) and the detector type installed at the beamline. The default attenuation corresponds to the current value at the beamline; usually, no attenuation is necessary for the test images, except for BL12-2. Changing the resolution will make the detector move at the start of the image collection. The Laue symmetry and cell can also be given if known; the multicrystal strategy mode can also be selected here; this will change the strategy program used to calculate the oscillation range for data collection from BEST or MOSFLM.

    For SAD and MAD experiments: The number of heavy atoms and the number of residues per monomer may be also specified here. If given, they will be used in the dose calculation for the strategy.

    Figure 11: Options for test data collection and autoindexing.
    Image autoindex-setupother2


Multicrystal strategy

The multicrystal strategy mode allows to select the optimal starting and final orientation phi in cases where a full data set cannot be collected from a single crystal, often because of radiation damage during data collection of microcrystals, metalloproteins or samples at room temperature. This option can be selected by choosing the multicrystal option in the Other Options setup page.

Upon selecting the multicrystal strategy option, the strategy calculation program will change to MOSFLM and a menu to select multicrystal strategy options will appear.

Figure 12: Options for multicrystal strategy.
Image autoindex-setupother-multi

If the crystal mounted is the first one in the multiple crystal experiment, the Previous run in the multicrystal series input box must be left empty. Once a strategy calculation run has been calculated for the crystal, a file containing information about the orientation of the crystal and the associated strategy is written out in the webice directory for the run. When another crystal in the multicrystal series is mounted, the immediately preceding run must be selected using the drop down menu. This will automatically read in and use the file storing the strategy parameters for all previous crystals in the series and updated with the values determined from the current sample.

The user can specify the expected rotation range that can be collected from the crystal before damage takes place in the rotation angle input box, which defaults to 10 degrees. For the first sample, accurate selection of the rotation range is not critical and could be even left blank -in this case, the program will calculate the standard range to maximize completeness for the crystal symmetry and orientation. However, in general, a fuller completeness will be achieved with a more accurate estimate of the rotation range, based either on the dose calculated by the strategy or the actual decay observed from previous samples.

When a previous run is selected, the program displays the starting and end phi angle determined by the strategy for the previous runs. If these values do not match the data collected from the crystal they can be edited to reflect that data that is actually available (for example, if the crystal was either damaged earlier or lasted longer than the input rotation range)

Note that, for multicrystal MAD and SAD experiments the input angle range will be used, unlike in single crystal mode, where twice as much data are collected by default to improve the multiplicity


Importing the screening strategy

If the sample has been mounted and analyzed during a screening experiment, a new run can be created to view the optimal data collection strategy for that sample. To do this from the autoindex tab, follow these steps:

  1. Make sure that the appropriate beamline is selected. Click on New Run and then on the Import button in the New Run Form page. The software will display some information pertaining to the sample. A default run name will be built from the spreadsheet ID and the sample name is also supplied.

  2. Edit the default run name if desired, and click Import Run.

Note: It is also possible to import the strategy from the Screening Tab or by launching the Web-Ice application from the screening tab in Blu-Ice.

Important: The import feature is only enabled if the sample is still mounted on the goniometer. If the sample has been dismounted, the autoindex must be repeated in order to obtain the optimal phi range for data collection.

Executing the run

After the run setup has been completed, a summary of the user's choices is displayed. The user can click the Edit Setup button to change the choices or the Run button to start collecting or autoindexing the images.

The Run Status will be periodically updated in the Run box. Messages will be displayed when autoindexing results are ready for viewing. It is possible to use other Web-Ice tools while the processing is running.

Clicking the Update button will reload the run contents.

Aborting autoindexing

The user may choose to abort the data collection or autoindex run by clicking the Abort button displayed after starting autoindexing.

Figure 13: Running Autoindex
Image autoindex-running


Viewing the run results

The Selected Run displays a series of links to facilitate viewing all the run analysis results.

Figure 14: Selected Run navigation links
Image autoindex-links

Scan

The [Scan] page only appears if the Perform Fluorescence Scan option was selected for a MAD/SAD experiment. This page will display a graph showing the raw fluorescence counts as the scan is being carried out, as well as messages from the scan operation. Once the scan is processed with autochooch, a plot of f' and f'' will also be available. The plots can be saved as PNG images by right-clicking on them and selecting the Save Image As... option from the menu.

Log

The [Log] link allows the user to inspect the detailed autoindexing log (with information about each program and version that is run and exit status) and data collection status on the connected beamline. The latter log is shared with Blu-Ice

Autoindex Summary

The [Autoindex Summary] page displays the refined values for the direct beam coordinates (Beam x and Beam y) and the distances; an initial estimate for the mosaicity based on how many of the observed spots are predicted by the best solution; and a list of indexing solutions.

For each indexing solution, the programs displays the following:

  • Metric fit: This is a measure of the distortion of the unit cell required to fit the symmetry. This increases with the symmetry but often remains below 0.1 for good solutions.

  • rmsd: Root mean square deviation between the observed and the predicted pattern. Usually less than 0.5 for good cases.

  • # spots: Number of spots predicted by the solution (up to a maximum of 600).

  • Crystal system, Unit Cell (in Å) , Volume (cell volume).

Solution

A summary of the refinement and integration results for the indexing solution determined by LABELIT is shown at the top of the page. Note that for strong diffraction images the estimated resolution limit may extend well beyond the limits of the detector; in order to collect data to this resolution, you should then move the detector closer to the sample; changing the exposure time by a large amount may have an effect on this estimate.

Note: be aware that the estimated mosaicity value can be underestimated for well diffracting crystals. The estimated value treated as suspect until several contiguous images are available for postrefinement.

The profile fitting for the central region of the images and intensity integration statistics are displayed below the ''Integrated Solution'' table.

To access the entire MOSFLM log and summary file for the integration, see the [Details] link.


Predictions

To view the predicted spots, click on the [Prediction] tab. This page displays the observed spots, and the predicted spots after autoindexing and integration with MOSFLM for either image used for the autoindexing. The image display parameters can be modified as described in the Image Viewer documentation.

The observed spots used for autoindexing are shadowed in green.

The predicted spots surrounded by a blue circle are fully recorded. The ones surrounded by a yellow circle are partially recorded.

Ice rings and the initial resolution estimate (before autoindexing) are also marked.


Details

[Details] displays the directory tree containing all the files generated by Web-Ice during the autoindex and integration, including scripts, orientation matrix files, logs and output reflection files. Access to these files is useful to validate the image analysis results and to refine the strategy calculation by making changes to the input files and re-running them manually. The contents of non binary files can be displayed by clicking on the file name.

  • The top directory in the tree /data/user-id/webice/autoindex/run-id contains

  • The LABELIT subdirectory contains the autoindexing results and the MOSFLM matrix, input scripts and output files for each possible autoindex solution

  • The solutionxx subdirectory contains the integration results for the solution number xx (since 2015, the results for the triclinic solution are not calculated). It also contains the dose calculation results (in the RADDOSE directory) and the strategy input scripts and log files for each possible symmetry consistent with the autoindexing solution (in directories named after the corresponding Laue group).

  • The SCAN directory contains all the output files from the fluorescence scan (if a MAD or SAD experiment has been selected). These files are also written to the image directory.

  • If MOSFLM is chosen as the strategy program, a BEST directory containing the input and log files from the program BEST (used to calculate the exposure time) will also be generated.

Strategy

The strategy calculation results are shown under the [Strategy] tab.

The optimal oscillation range for the Laue group determined by LABELIT and the overall completeness are displayed in the ''Space Group'' table. An upper limit for the oscillation per image (delta phi) is also shown in this table. Above this delta phi value, more than 3% reflection in one or more images in the phi range will overlap. Note: The maximum delta value is based on the mosaicity estimation. If the mosaicity is underestimated, the delta phi value will be overestimated.

Figure 15: Space group summary table. Unless the space group was preselected before autoindexing, the optimal collection range is calculated for all the lowest symmetry group in the predicted lattice.
Image autoindex-green-arrow

The full recommended data collection strategy for the determined space group is displayed at the bottom of the page. The BEST strategy calculation results can also be displayed by selecting the appropriate tabs above the data collection strategy table1. Clicking on the phi ranges displayed in the table will toggle between the strategy to optimize anomalous data and unique data completeness.

Figure 16: Strategy navigation tabs to toggle between the strategy summary table and the completeness and overlap analysis for each Laue class in the autoindexing solution.
Image autoindex-strategy-links

The Data Collection Strategy table shows all the required input for a Blu-Ice data collection run based on the strategy calculation results. Below is an explanation of how each quantity is determined. If the results are not suitable for the users' experiment it is possible to re-run the calculations as explained in the troubleshooting section.

  • Experiment type: This will also display a warning if a non-default total oscillation range is selected in the Space Group Table (e.g., if the user selected a SAD experiment, but clicked on the phi range maximizing unique completeness).

  • Mosaicity and Score: These values are shown for information purposes only. If the mosaicity is flagged as poor, or the score is low, it is strongly recommended to inspect the images before proceeding to data collection.

  • Oscillation start and oscillation end: For native data collection, the default values optimize the unique data completeness.

    For SAD and MAD experiments, the values maximizing the anomalous pair completeness are selected instead, and the total oscillation range is doubled (ie, the multiplicity of the data is doubled for each wavelength), since increased multiplicity is better for anomalous signal detection.

  • Oscillation angle: The oscillation angle calculated by BEST is used. This should not exceed the maximum delta phi value displayed in the strategy summary table.

  • Oscillation wedge: For conventional data collection, it is suggested to collect the entire run in a single continuous patch. For MAD and SAD experiments, data collection in wedges is preferable, to minimize the amount of damage between Bijvoet or dispersive pairs. When feasible, the wedge size is chosen so that the total dose received by each wedge of data does not exceed 0.025 MGy. However, if the dose wedge-based size is smaller than a minimum wedge size (based on the amount of time needed to change energy on each individual beamline), the latter will be selected instead.

  • Resolution: This is the resolution for which the strategy has been calculated. A warning is displayed if the maximum resolution predicted from the image integration statistics (shown under the [Autoindex] and [Solutions] links) is higher than the value at the detector edge.

  • Exposure time: BEST is used to estimate the optimal exposure time to collect data with an $I/\sigma I$ of 2 at the resolution given above. For SAD and MAD experiments, the $I/\sigma I$ target is chosen so that a default signal of 3% can in principle be measured at low resolution. If the user provides the number of anomalous scatterers in the molecule, the signal will be estimated from that amount. Note: While this is a better criterion to determine the strategy in phasing experiments, theory and reality may differ. If the expected anomalous signal is small, increasing the data multiplicity beyond what is recommended by the software may be a good course of action.

    If the exposure per image is too short, it will be reset to the minimum allowed value; in this case, a suitable beam attenuation will be calculated to avoid image overexposure.

  • Attenuation; Calculated to avoid too short exposure times2.

  • Detector to sample distance: Detector to sample distance to obtain the resolution given above. If the predicted resolution is shorter, a warning will be displayed with the required sample-to-detector distance.

    Note: If the detector distance to achieve the maximum resolution in the warning differs noticeably from the other value, the values for the exposure time and oscillation angle may not be valid any more. In this case, it is recommended to recollect test images at the new distance and repeat the strategy calculation.

  • Beamstop Distance: Minimum beamstop to sample distance required to collect data at a low resolution of 40 Å. To collect data at a lower resolution (at the expense of a higher image background) the beamstop must be moved further back. Note that the optimal value is different at wavelengths other than the one used to collect the test images.

  • Energy: For native crystal experiments, the energy is read from the image header; this is also the default for SAD experiments, unless the user has selected the element from the periodic table with the Perform Fluorescence Scan option. In these cases, the software may display a warning suggesting a higher energy for crystals diffracting to very high resolution; if so, it is recommended to recollect test images with appropriate values of energy and detector distance and repeat the strategy calculation.

    In all other cases, the optimal energies for MAD and SAD data collection will be calculated based on a fluorescence scan or values supplied by the user.

  • Detector type: This is read from the image header and displayed for information purposes.

  • Detector mode: Dezingering is suggested for CCD detectors if the exposure time per image is longer than 30 seconds.

  • Inverse beam: This option is selected for SAD experiments (inverse beam = Yes). For the other experiment types this is set to No (no inverse beam).

  • Number of images: Obtained by dividing the total oscillation range by the oscillation per image.

  • Absorbed dose: Estimate of the absorbed dose assuming no heavy atoms present in the sample, a solvent content of 50% and a uniform crystal size of 100 microns. For MAD experiments, the dose at each energy will be given. If the total dose exceeds the Garman limit, a warning will be displayed; for MAD and SAD experiments, a more strict limit (half the Garman limit) is used, as the accurate measurement of anomalous differences is easily affected by radiation damage; in addition, for MAD experiments, only data at two energies will be collected. Other measures to reduce the absorbed dose are left to the user's discretion: decreasing the exposure time per image is the best option in most cases. However, if an abundant supply of crystals is available, decreasing the number of images is also an option.

    Important: The dose calculated for MAD and SAD experiments can be significantly underestimated if the protein buffer/cryoprotectant contains a high concentration of ions such as Ca, Cl or SO4 or heavier atoms). Please see the troubleshooting section to find out how to calculate a more realistic dose in this case.

Initiating data collection

The values given in the Strategy table can be used for subsequent data collection using the Export Run Definition, Recollect Test Images and Collect Data set buttons in the [Strategy] page.

Figure 17: Options after strategy calculation.
Image autoindex-strategy-buttons

  • Export Run Definition: This button does not initiate data collection automatically, but fills out a Blu-Ice run entry. The user can then open a Blu-Ice client and start the data collection. This option is useful if the user wishes to modify certain beamline operation parameters which are not accessible from Web-Ice )(e.g., dose mode or beam size) or to queue the data collection run. Some of the settings (oscillation start and end, oscillation angle, exposure time, detector and beamstop-to-sample distance and energy) can be edited after clicking the button and before the run is exported.

  • Recollect Test Images: This is recommended if the user wishes to repeat autoindexing and strategy calculation based on the warnings given by the strategy. Clicking the button allows the user to edit the oscillation angle, exposure time, detector and beamstop distance and energy before recollecting two images and repeating the analysis.

  • Collect Data Set: This button starts data collection. The user may edit the oscillation start and end, oscillation angle, exposure time, detector and beamstop-to-sample distance and energy before data collection is initiated.

Important: Web-Ice will recalculate the dose when the exposure time, attenuation or oscillation range are changed in the Export/Collect page. However, the resolution does not change. The resolution corresponds to the diffraction limit of the crystal, and it is difficult to predict the new resolution when the exposure is changed without analyzing another pair of test images first.


Footnotes

...strategy-links1
Before September 2015, the software calculated the strategy for all the space groups; this took a long time and has been discontinued.
... times2
Note that the actual attenuation may differ slightly from the calculated attenuation. Also, on MAD experiments, the attenuation at the remote wavelength will be slightly different to the attenuation at the edge energies.