Project

General

Profile

Forums » Leginon »

semi-automated data collection for screening negative-stain grids

Added by Junjie Zhang over 10 years ago

Hi,
Is there any application in leginon that is customized for semi-automated data collection to screen a negative-stain grid. These MSI applications work great when there is a very nice grid. But sometimes, the quality of the stain may not be perfect through out the grid and one has to check at a higher mag instead of the grid preset.
Is there a way we can just
(1) quickly check each grid square on the fluorescent screen and mark down the X,Y,Z coordinates of those good squares. One can switch to high mag to see the particle stains.
(2) manually or use Raster to queue the local target spots on each square for z-focus, focus and exposures.
(3) submit all these targets and perform automation.
Thank you!


Replies (4)

RE: semi-automated data collection for screening negative-stain grids - Added by Junjie Zhang over 10 years ago

Shall I use the following procedure to perform data collection as described above?
(1) use MSI-raster application.
(2) generate the atlas of the entire grid (about 20 mins) with the gr preset.
(3) in square targeting, do not click on the squares since I can not judge the stain quality from looking at the atlas of the grid. Instead, I send the ef preset to the microscope and manually move the stage. I then visually inspect the stain quality with the ef preset. Once I find a good square, I refresh the current position (orange cross hair) on the atlas shown in leginon and put a green cross hair in the same position for later use.
(4) repeat step 3 until enough squares have been identified.
(5) submit these squares for subsquare targeting with raster by define the spacing of each exposure.
(6) in each subsquare just do one exposure.

Is there a better way for doing step 3?

Thanks!

RE: semi-automated data collection for screening negative-stain grids - Added by Anchi Cheng over 10 years ago

Here is one answer from users in the group.

I do exactly what he does. I always look at each square first because if I'm about to collect 500+ images on the same square i think it's worth the ~10 second visual check on the screen to make sure it looks good.

If I were to do this, I would do it a bit differently in step 3. I would use the atlas to pick a candidate square from the atlas and submit it to get a square image and be at the general area with reasonable stain. I would have "Subsquare Targeting" in user verification mode and in queuing. If the square is good using ef as in your protocol, I will keep
the automatic raster targets, submit them to queue. If the square is bad, then I use "control-shift + mouse right click" to remove the targets and submit the empty target list to queue. This way, I know where I have been.

With queuing on, you can also skip atlas, and start imaging a good sq mag with simulated target tool. This means you would save the good positions in TUI or in Leginon Naviation node while you screen for good square. Again, put raster targets you want in each square in queue.

One caution note, target selection on a square image will be imaged at the Z of the square image unless you include a focus target which adjust the stage Z value.

    (1-4/4)