Tips and Tricks
From NE-CAT Public Wiki
Contents |
Shipping
1. Make sure the pin length is uniform and you pack one extra empty pin for alignment
2. Dry shipper should be filled with LN2 at least two days in advance to properly cool it down
3. Shipping delays DO occur - do NOT send your dewars on Thursday for Friday delivery - it's too risky.
4. Always cover the crystals canes with plastic sleeves ( Hampton Research ) - to prevent cups to fall out of canes because of carrier mishandling.
5. Learn what to expect for data collection (unit cell, symmetry, resolution), if you are collecting data for your collegues.
6. Prepare the hard disks for your data backup : remove old data and create an ext3 (linux) partition; on average 300-400 Gb is needed for one day of data collection
7. Pack your favourite tools for crystal handling and mounting ( no kappa arks are allowed)
Arriving at APS
1. Arrive on day before your start time : you will have a good sleep before the hectic day of data collection; besides travel delays may occur and several groups lost their time.
2. Finish all mandatory APS safety training before coming to the beamline
3. Initial sector orientation/refreshement is mandatory for ALL members of your group, even if it is a split group. Watch for couple hours over the shoulder of your collegues collecting data together with the beamline staff, that experience will be valuable when you will bealone at night.
4. Did you lose your car key? Or is there flat tire? Seek help from the Floor coordinator by dialing 2-0101.
Data collection
1. Alignment of the crystal with the beam is very important for the proper scaling. The large amplitude waves of chi2 for x- and y-positional refinement during the integration is an indication of the misalignment ( for non plate-like or needle like crystals) . Please check the beamline manual how to repeat the beam alignment.
2. Always collect first snapshot with plenty of Al filters - as a rule of thumb: use 1.5mm for big crystals, 1mm for intermediate (~0.1x0.1x0.1mm) and 0.5 mm Al for a small crystals. Check the maximum intensity in the ADX graphics windows and adjust attenuators to reach 40000-60000 counts. Use the program RADDOSE to estimate the possible number of diffraction images at the current energy , for a specific crystal and attenuators to reach the absorbed dose of 1Gr.
3. High mosaicity, splitted spots on a first snapshot - thinking about mounting of a new crystal - wait: Try doing simple annealing. Either block the cold stream for 10-20 seconds with your credit card or put the crystal back into mother liquid and freeze it again. You have nothing to lose, but in many cases after annealing the diffraction is improving.
4. Nice, promising diffraction at 0 and 90 degrees: do not rush to data collection windows. Process your shapshots (either by HKL2000 or mosflm), determine space group and run the STRATEGY. Starting your data collection at the right angle will provide you with more complete data and higher redunduncy. Collect ~50% more oscillation range, as suggested by startegy routine, to decrease the detrimal effect of using higher crystal symmetry or to estimate the real radiation damage for next crystal data collection startegy.
5. Do a real time data integration and scaling: many problems may be intercepted while the crystal is still mounted on the goniometer. Learn the effect of the radiation damage on your crystal. Use the results of scaling to optimize the data collection startegy for the next crystal. Running XTRIAGE with new scalepack.sca file will help.
6. Are you planning to check a new protein crystal for the first time? Always bring at least 3 or 4 crystals mounted in the capillary. The capillary mounted crystals will shows the "base diffraction" of your crystal. You can use this as a reference to check whether the cryoprotectant spoils the diffraction quality of your crystal or not.
The best data collection strategy for small crystals
There are two approaches to collect good data from small crystals on MD-2.
1. This one is fast and simple, but not fully optimal.
Let assume you have elongated rod 20x20x100 micron, oriented close to parallel to rotation axis. The calculated in HKL2000 strategy suggest to collect 120 degrees sweep.
Now you collect ~100 images (100 deg.) from one portion of the crystal with low flux: index, integrate and scale all frames. Check the scale and B-factor as a function of frame number : you may safely use only frames with B-factor less than 10-15 A2. For example, only first 30 frames have B factor less than 10 and reasonable chi-square.
Based on previously collected strategy, you need to collect data only from four (4x30=120deg) fresh portions of the crystal, that is feasible with your 100 micron long crystal. Visually divide your crystal into four or five (better) portions for data collection.
The next step is to move and align the new fresh portion of the crystal. Collect another 30 frames, starting phi angle should be deduced from the first run (end of 30th frame). Repeat the same procedure two more times , so 120 degrees sweep is collected from your crystal.
Process (integrate and scale) all data in one HKL2000 session and analyze the results: collect some extra data from fifth fresh portion of the crystal, if warranted, by scaling results. This approach is based on the assumption that it's possible to collect several small datasets from one crystal if not, you should use the second protocol:
2. This is an advanced calculation of the optimal data collection strategy.
Start again by collection ~100 images from one portion of the crystal with low flux: index, integrate and scale all frames. Check the scale and B-factor as a function of frame number : you may safely use only frames with B-factor less than 10-15 A2. For example, only first 40 frames have B factor less than 15 and reasonable chi-square and your crystal is so small that only one more fresh portion may be used for data collection. Recalculate all scaling for 40 frames only, so the corresponding "output.sca" with first good data is in your working directory.
Now we will calculated the strategy for the next data collection step: copy the file "/home/necat/Xtal_Soft/strategy/strategy.in" into your home directory, and start editing it. This file already contains some explanations for the parameters to be used (any line with # is ignored). The most important are the options to include data already collected from one (or more) previous crystal(s) and the options to collect data in discontinuous segments.
The minimum input file is the following:
XFILE ./your_imagename_001.x
SPACE GROUP C2221
ALREADY MEASURED ./output.sca
PSFILE stat.ps
Run /home/necat/Xtal_Soft/strategy < strategy.in and analyze the text output.
By default the software will calculate the best starting angle and total oscillation range to reach 95, 97, 99 and 100% completeness. The file stat.ps contains all results in a visual form (use ggv stat.ps to view it).
If no results are produced, it is worth to include the line: COMPLETENESS X1 X2 ... XN (like 70, 80, 90). The program will use the values X1 ... XN, as the desired completeness rather than its default values (95, 97, 99 and 100%). If planning for SAD data collection , you need to include ANOMALOUS flag.
It is a good practice to start the next data collection sweep a few degrees below the predicted one to ensure that all the partially recorded reflections had been recorded fully. Collect the next sweep of data, process and scale all data: the above procedure may be repeated iteratively until you have 100% complete data.
Word of caution: It is very easy to burn the protein micro crystal on our beamline. It is imperative to use the low flux to reach close to 100% completeness, even at the expense of resolution. Remember, a 99% complete data at 3.2A resolution will be perfect for easy structure solution and refinement as compared to 70% complete data at 3.0A.
If this approach was succesfull, please cite: Raimond B.G. RAVELLI, Robert M. SWEET, John M. SKINNER, Albert J.M. DUISENBERG and Jan KROON : "STRATEGY: a program to optimize the starting spindle angle and scan range for X-ray data collection" J. Appl. Cryst. (1997). 30, 551-554
Collecting an EXAFS Scan
To collect an EXAFS scan from your protein crystal use the biggest crystal for EXAFS and save the best crystal for diffraction data collection. The crystal for the EXAFS scan needs not be single, but the size matters. Currently, the beam flux used to collect EXAFS is very high, as compared with previous years of ID-C operations, and it WILL damage your crystal.
1. Change wavelength to the desired edge you will be scanning across.
2. Choose “BEAM LOCATION” on the MD2, and center the beam.
3. Choose “DATA COLLECTION” on the MD2.
4. In console script, click on “SCINT DET POS” in order to move the fluorescence detector close to your sample.
5. Adjust attenuation, so that counts on the DT SERVER (rightmost computer screen) are between 2 and 5 when the shutter is open.
- Near 12658 eV, 1250 μm of attenuation is a good starting point.
- Near 8000 eV, 250 μm of attenuation is a good starting point.
6. Click on “EXAFS SCAN” in the console script.
7. Click on “Browse” and select the file that corresponds to the element you are scanning.
8. Put a filename in the OUTPUT field with the following format: <*Element*>_<*Crystal Description*>.dat
9. Press “ACCEPT RUN SCAN” and “EXIT XAFS CONFIG” to run the scan or “ABORT XAFS SCAN” to abort the scan.
10. Allow the scan to run (~5 minutes for a 40 eV scan). The raw data will be plotted on the rightmost computer screen during the scan.
11. Once you hear “EXAFS energy scan complete.” (the second of two audio prompts), you may run chooch, change your sample and/or change energy.
12. You can close the scan plot on the rightmost computer screen by selecting “MASTER OPS” and then “CLEAR MASTER EXAFS NTGRAPH”.
NOTE: For a more detailed description of how to collect an EXAFS scan, see p. 16 of the 24-ID-C Operational Manual (http://lilith.nec.aps.anl.gov/userprogram/24ID_manual.pdf)
Hard Drives
Formatting
We recommend that all portable hard drives be in ext3 format. We cannot write to NTFS or Mac formatted hard drives. We have experienced problems with FAT32 formatted drives. After the disk exceeds 60% full, data transfer slows and eventually halts, preventing users from fully backing up their data. Only use FAT32 if you not planning to collect very much data or you are not intending to take your data home.
Seagate FreeAgent Drives
These drives have known problems interacting with Linux. There is, however, a known fix that will prevent the drives from sleeping. The command is as follows for a drive mounted on /dev/sda: sdparm --clear STANDBY -6 /dev/sda Of course, to run sdparm, you must be root.
