[3dem] Faulty autogrid rings - (Re: image rotation on grid maps)
Andrea Nans
andrea.nans at gmail.com
Thu Oct 8 07:00:40 PDT 2020
Hi there,
We had some faulty autogrids from an order placed in January 2020 and again
in August 2020. The proportion of bad rings was on the lower side, perhaps
around 1 in 10.
Andrea
Francis Crick Institute, London.
On Thu, Oct 8, 2020 at 2:54 PM Baxa, Ulrich (NIH/NCI) [C] <
baxau at mail.nih.gov> wrote:
> Hi Duncan,
> We have received a shipment of C-rings from TFS in April 2020 and many of
> the bags had faulty rings (some of them were almost 100% bad). So we can
> confirm that the problem existed at least since April 2020.
> Best
> Ulrich
>
> -------------------------------------------------------------
> Ulrich Baxa
> National Cryo-EM Facility
> Cancer Research Technology Program
> Leidos Biomedical Research, Inc.
> Frederick National Laboratory for Cancer Research
> P.O. Box B
> Frederick, MD 21702
> Phone: 301-846-1559
> baxau at mail.nih.gov
>
> web:
> https://urldefense.com/v3/__https://www.cancer.gov/research/resources/cryoem__;!!Mih3wA!QJzFSR5UIfWg-Wy7jt7rzviSlwKxfZcDknnBc_dfO-QBM4A94oCApsAbilskFg4WSA$
>
>
>
> On 10/7/20, 10:57 PM, "Mike Strauss" <mike.strauss at mcgill.ca> wrote:
>
> Hi Duncan,
>
> to my knowledge they don’t have lot numbers. Thermo says they’re
> looking into it, but haven’t made any conclusive statements. In Canada, we
> are being assured that we will get replacements for the faulty grids, but
> on some bags it’s over 60% failure.
>
> Best,
> mike
> --------
> Prof. Mike Strauss
> Anatomy and Cell Biology / FEMR
> McGill University
> (514) 398-8084
> mike.strauss at mcgill.ca
>
> > On 7. Oct 2020, at 21:45, duncan.sousa at rcn.com wrote:
> >
> > Hi,
> >
> > Does anyone have lot #'s from TFS for the faulty autogrid rings?
> We're recently had more bad rings delivered to us out here at Johns
> Hopkins. This problem appears to have first been noticed in late July /
> early August.
> >
> > I've attached images from an analysis done by Joeseph Felt (HRMEM
> facility at UBC in British Columbia, Canada).
> > Bad URL Removed - see why -
> https://urldefense.com/v3/__https://ees.sps.nih.gov/services/Pages/Anti-Virus.aspx?A2=ind2008&L=CCPEM&O=D&P=94300__;!!Mih3wA!QJzFSR5UIfWg-Wy7jt7rzviSlwKxfZcDknnBc_dfO-QBM4A94oCApsAbilvXwv-6RA$
> >
> > Kind Regards,
> > Duncan
> >
> > Duncan Sousa, PhD
> > Johns Hopkins University
> > School of Medicine
> > 855 N Wolfe St., Rangos 432
> > Baltimore, MD 21205-2185
> > 1-410-955-3750
> >
> >
> > From: "Craig Yoshioka" <craigyk at gmail.com>
> > To: "Joeseph Felt" <joesephfelt at gmail.com>
> > Cc: "Mike Strauss" <mike.strauss at mcgill.ca>, "David Mastronarde" <
> mast at colorado.edu>, serialem at colorado.edu, "Chen Xu" <Chen.Xu at umassmed.edu>,
> "Florian Rossmann" <frossman at mail.ubc.ca>, "Alison Berezuk" <
> alison.berezuk at ubc.ca>
> > Sent: Monday, August 24, 2020 2:45:43 PM
> > Subject: Re: image rotation on grid maps
> >
> > Hi Joseph,
> >
> > Ridge also confirmed on the bad grids here. Still seeing if our
> users can chase down a lot #.
> >
> > One weird thing is I didn’t notice an obvious phase transition in
> the ice, it did seem to be mostly sublimating from the surface… wonder if
> this could be designed to be a controlled process…
> >
> > I’ve told TFS before they should add a sonicare toothbrush motor to
> the next Autoloader revision for knocking surface contamination off grids-
> I’ve had some serendipitous observations of this occurring, and it would be
> great to make it a controllable, activatable option, like sensor cleaning
> on DSLRs.
> >
> > Cheers,
> > -Craig
> >
> >
> > On Aug 24, 2020, at 11:16 AM, Joeseph Felt <joesephfelt at gmail.com>
> wrote:
> >
> > My thoughts exactly. I also plan on disassembling our recently
> clipped grids to compare the faulty ones with the ones that remained
> vitrified.
> >
> > All the best,
> >
> > Joeseph
> >
> > On Mon, Aug 24, 2020 at 11:11 AM Craig Yoshioka <craigyk at gmail.com>
> wrote:
> > The ridge is likely preventing the ring from fully expanding, and as
> a result pressing itself down onto the grid. I will see ask to have our
> faulty grids checked for this ridge to confirm as well.
> >
> >
> >
> > On Aug 24, 2020, at 10:45 AM, Joeseph Felt <joesephfelt at gmail.com>
> wrote:
> >
> > Hello all,
> >
> > Here are some images of the differences I'm seeing on the autogrid
> rings. The ridges along the inner wall are more obvious under our
> dissection microscope, but unfortunately it doesn't have a camera, so I did
> my best to take pictures comparing a ridged ring and a smooth ring through
> the eyepiece. I've also included images taken on our crystal screening
> microscope, but it was more difficult getting the angle right. I've
> labelled the rings as either ridged (R) or smooth (S).
> >
> > All the best,
> >
> > Joeseph
> >
> > On Mon, Aug 24, 2020 at 9:15 AM Joeseph Felt <joesephfelt at gmail.com>
> wrote:
> > Further testing with clipping at room temperature suggests that it
> was the reverse of what I had expected: the autogrid rings that are smooth
> on the inside have C-clips that are tight and do not shift, and the rings
> with grooves/striations along the inner wall had loose-fitting C-clips. I
> tested four of each, and it was consistent in those that I tested. I'll try
> to get some images and do more testing at cryo temperatures etc.
> >
> > All the best,
> >
> > Joeseph
> >
> > On Mon, Aug 24, 2020 at 8:45 AM Joeseph Felt <joesephfelt at gmail.com>
> wrote:
> > Hello all,
> >
> > This is Joeseph from the HRMEM facility at UBC. I've had a chance to
> look at the individual autogrid rings we have in more detail under a light
> microscope. The batch that has given us devitrified/rotating grids might
> have defects on the inside of the autogrid rings. I'm noticing that about
> half of the rings have little slots or grooves along the inner walls where
> the C-clip is presumably meant to sit, but the other half are missing these
> grooves, so it is just smooth along the inner wall. This may explain why
> the C-clips are shifting freely in the autogrid assembly. I'm going to try
> assembling some of them with the C-clips to confirm my suspicions.
> >
> > All the best,
> >
> > Joeseph
> >
> > On Mon, Aug 24, 2020 at 8:10 AM Mike Strauss <mike.strauss at mcgill.ca>
> wrote:
> > Hi Craig,
> >
> > including you, I now have word of 5 sites with similar issues
> appearing recently. Careful work by the group at UBC seems to show that
> the clipped grids are not secured properly - indeed they seem quite loose.
> This would explain the devitrification, and the rotation. So, either they
> are shipping out bad lots of cartridges, or clip rings (or both I
> suppose). We’re still working on tracking down the lot numbers.
> >
> > Do you have the lot numbers of your recent consumables?
> >
> > Regards,
> > mike
> >
> > --------
> > Prof. Mike Strauss
> > Anatomy and Cell Biology / FEMR
> > McGill University
> > (514) 398-8084
> > mike.strauss at mcgill.ca
> >
> > > On 17. Aug 2020, at 20:43, Mike Strauss <mike.strauss at mcgill.ca>
> wrote:
> > >
> > > Hi Craig,
> > >
> > > the short answer is, no, I didn’t get it resolved. Engineers
> working on the scope and my own vacation plans hampered an in depth
> investigation.
> > >
> > > Anecdotally, though, the affected grids did appear dry - perhaps
> sublimated... I haven’t looked at that with rigour, though, so it’s
> relegated to the realm of possibility for now.
> > >
> > > Best,
> > > mike
> > > --------
> > > Prof. Mike Strauss
> > > Anatomy and Cell Biology / FEMR
> > > McGill University
> > > (514) 398-8084
> > > mike.strauss at mcgill.ca
> > >
> > >> On 17. Aug 2020, at 20:35, Xu, Chen <Chen.Xu at umassmed.edu> wrote:
> > >>
> > >> This is interesting and possible. We had a single case that
> everything seems to be good but all of sudden ice is all gone.
> > >>
> > >> I suppose it is possible that forks have lose contact to grid in
> some cases. This is something we ought to keep close eye on. I wonder if it
> is possible for forks to build up dirt after some time in use.
> > >>
> > >> Best,
> > >> -Chen
> > >>
> > >>> On Aug 17, 2020, at 7:55 PM, Craig Yoshioka <craigyk at gmail.com>
> wrote:
> > >>>
> > >>> My working theory was that the clip rings are just a tad small
> and not being gripped by the forks. Backing this up was that it looked
> like the grid was drying out sublimating. I re-loaded the CG and it’s
> smooth as butter- so my current conclusion is that our users had a bad
> batch of clip-rings (we checked two of their grids and they behaved like
> this). Maybe Mike got the same bad batch….
> > >>>
> > >>>
> > >>>
> > >>>> On Aug 17, 2020, at 4:37 PM, Xu, Chen <Chen.Xu at umassmed.edu>
> wrote:
> > >>>>
> > >>>> Hi Craig -
> > >>>>
> > >>>> I would suspect it is more likely due to clipping not being
> properly tight. I never experienced this on an autoloader, but I did see
> similar on old FEI rotation, single-sided room temperature holder which
> uses c-clip ring too.
> > >>>>
> > >>>> Best,
> > >>>> -Chen
> > >>>>
> > >>>>> On Aug 17, 2020, at 7:09 PM, Craig Yoshioka <craigyk at gmail.com>
> wrote:
> > >>>>>
> > >>>>> Hi Mike,
> > >>>>>
> > >>>>> Did you ever diagnose this issue? The reason I’m asking, of
> course, is because we just saw some lmms like this ourselves:
> > >>>>>
> > >>>>> I think it’s a stage issue myself from watching it in
> continuous mode, the grid is jumping around quite a lot as the stage
> moves. Weird coincidence though, first time seeing this on a stage that’s
> been doing fine for 5+ years.
> > >>>>>
> > >>>>> <Screen Shot 2020-08-17 at 3.57.26 PM.png>
> > >>>>>
> > >>>>>
> > >>>>>> On Aug 10, 2020, at 9:27 AM, Mike Strauss <
> mike.strauss at mcgill.ca> wrote:
> > >>>>>>
> > >>>>>> Hi David,
> > >>>>>>
> > >>>>>> That’s what was strange. This was the 3rd atlas, where the
> first two were fine. The duration between the start of the atlas and the
> middle of it is about 3 min. But the first image of the first atlas had
> been taken 20 min previously.
> > >>>>>>
> > >>>>>> I’ll first deactivate the beta tilt and see if the problem
> persists.
> > >>>>>> I’ll then check the LowMagNormalizationDelay setting, and the
> behaviour by script after this data collection is through. I’ll also test
> for image rotation as a function of defocus in LM. I did notice that some
> of these images may not be the same size (full frame), so I was wondering
> if there is an imposed rotation somehow (perhaps the one Craig mentioned
> earlier). But I still can’t explain the change in rotation.
> > >>>>>>
> > >>>>>> I will let you know what I find.
> > >>>>>>
> > >>>>>> Best,
> > >>>>>> mike
> > >>>>>> --------
> > >>>>>> Prof. Mike Strauss
> > >>>>>> Anatomy and Cell Biology / FEMR
> > >>>>>> McGill University
> > >>>>>> (514) 398-8084
> > >>>>>> mike.strauss at mcgill.ca
> > >>>>>>
> > >>>>>>> On 10. Aug 2020, at 10:25, David Mastronarde <
> mast at colorado.edu> wrote:
> > >>>>>>>
> > >>>>>>> Hi Mike,
> > >>>>>>>
> > >>>>>>> I confess I did not grasp the point of your original
> images. Now I see it: the left size is very rotated, the second column is
> less, by the middle things look normal.
> > >>>>>>>
> > >>>>>>> I agree there is no reason for the actual focus to change
> during the montage. Now what comes to mind is some tremendously slow
> recovery from normalization or some other transient, such as the focus
> change itself. At the start of the montage (bottom left) it is the most
> extreme, and it recovers by the third column.
> > >>>>>>>
> > >>>>>>> What is the time duration of this portion of the acquisition?
> > >>>>>>>
> > >>>>>>> Do you have a setting for LowMagNormalizationDelay?
> > >>>>>>>
> > >>>>>>> Perhaps you can explore this with a script that goes into
> LM, sets a defocus, and takes a few single images with delays between them.
> > >>>>>>>
> > >>>>>>> David
> > >>>>>>>
> > >>>>>>> On 8/10/20 6:40 AM, Mike Strauss wrote:
> > >>>>>>>> Hi David,
> > >>>>>>>> View and search montages are fine. Just LM. I’m not sure
> why the defocus is set so high. By the same token, why would it change
> throughout the montage?
> > >>>>>>>>
> > >>>>>>>> mike
> > >>>>>>>>
> > >>>>>>>> -----
> > >>>>>>>> mike strauss
> > >>>>>>>> miPhone
> > >>>>>>>>
> > >>>>>>>>> On Aug 10, 2020, at 1:07 AM, David Mastronarde <
> mast at colorado.edu> wrote:
> > >>>>>>>>>
> > >>>>>>>>>
> > >>>>>>>>> Hi Craig,
> > >>>>>>>>>
> > >>>>>>>>> Oops. It looks like LM is excluded in only 2 of 6 places
> that call for a defocus-adjusted stage matrix, and montaging is not one of
> them.
> > >>>>>>>>>
> > >>>>>>>>> I think of LM as having a pretty big dependence on
> defocus, relative to the range of defocus that gets applied, but I guess
> that per micron it is a smaller effect than in nonLM. In any case, there's
> no reason the dependence should be the same, and the adjustment should not
> be applied in LM. It's best to settle on a standard focus for LM, calibrate
> the stage shift there, and take montages there. But does that happen when
> doing View or Search montages? As I said, I'm not sure.
> > >>>>>>>>>
> > >>>>>>>>> The adjustment does not extrapolate, so for the kinds of
> LM defocus we are talking about here, there wouldn't be any variation in
> the (inappropriate) adjustment ofr defocus.
> > >>>>>>>>>
> > >>>>>>>>> David
> > >>>>>>>>>
> > >>>>>>>>> On 8/9/20 10:20 PM, Craig Yoshioka wrote:
> > >>>>>>>>>>
> > >>>>>>>>>> There is also the rotation/scale calibration for large
> defocus changes, I assumed SerialEM would not apply this correction in LM
> (where the image rotation and magnification change very little with focus
> relative to M/SA), but I’ve never checked.
> > >>>>>>>>>>
> > >>>>>>>>>>
> > >>>>>>>>>>> On Aug 9, 2020, at 4:28 PM, David Mastronarde <
> mast at colorado.edu> wrote:
> > >>>>>>>>>>>
> > >>>>>>>>>>> Hi Mike,
> > >>>>>>>>>>>
> > >>>>>>>>>>> I agree that this looks like a defocus problem - unless
> you tell us the defocus was the same. If you turn off "Align pieces in
> overview", it will be easier to see the pattern of results and how much the
> stage calibration is off.
> > >>>>>>>>>>>
> > >>>>>>>>>>> LM montages were originally set up to be taken at a
> standard focus (whatever the stage calibration was done at, if I recall).
> But it's not clear how this interacts with using View mode with an imposed
> defocus offset.
> > >>>>>>>>>>>
> > >>>>>>>>>>> David
> > >>>>>>>>>>>
> > >>>>>>>>>>> On 8/9/20 4:49 PM, yanxiang cui wrote:
> > >>>>>>>>>>>> Hi Mike,
> > >>>>>>>>>>>>
> > >>>>>>>>>>>> From the first image, the defocus shows -3.41mm, which
> is already very big.
> > >>>>>>>>>>>>
> > >>>>>>>>>>>> What's the defocus you used in the second map you
> showed in this email ? I normally set the defocus of view mode to
> -0.8mm.
> > >>>>>>>>>>>>
> > >>>>>>>>>>>> I have some experience that if that defocus is too far
> away from eucentric focus, the polygon maps will be very strange.
> > >>>>>>>>>>>>
> > >>>>>>>>>>>> You can have a check if this is the reason.
> > >>>>>>>>>>>>
> > >>>>>>>>>>>> Best regards
> > >>>>>>>>>>>> Yanxiang
> > >>>>>>>>>>>>
> > >>>>>>>>>>>>
> > >>>>>>>>>>>>
> > >>>>>>>>>>>> On Sun, Aug 9, 2020 at 9:39 AM Mike Strauss <
> mike.strauss at mcgill.ca> wrote:
> > >>>>>>>>>>>> Hi,
> > >>>>>>>>>>>> We have a strange problem on our Krios running SerialEM
> 3.8.0 beta (Feb4,2020) and using a quantum K3.
> > >>>>>>>>>>>>
> > >>>>>>>>>>>> Periodically we have an apparent in-plane rotation of
> our grid when we’re acquiring grid maps. The stage calibration seems to be
> pretty good, and we often get good results (see first screen shot), but in
> the last few weeks, portions of the grid map will have a rotation (second
> screen shot). You can see this as the grid bars run in different
> directions. The grids are flat, and when we reimage them, the resulting
> montage may be totally fine, or may have rotated images in other parts of
> the grid. In some cases the rotation can be up to 30 degrees.
> > >>>>>>>>>>>>
> > >>>>>>>>>>>> I’ve checked the beta tilt, and that seems to be stable
> - at least the encoders don’t show it moving. The magnification of the
> image does not appear to be significantly different And the microscope
> behaves reasonably normally in other modes.
> > >>>>>>>>>>>>
> > >>>>>>>>>>>> Does anyone have suggestions why this may be occurring?
> This throws off our ability to navigate to grid squares reliably, and it’s
> quite frustrating.
> > >>>>>>>>>>>>
> > >>>>>>>>>>>> Regards,
> > >>>>>>>>>>>> mike
> > >>>>>>>>>>>>
> > >>>>>>>>>>>>
> > >>>>>>>>>>>> --------
> > >>>>>>>>>>>> Prof. Mike Strauss
> > >>>>>>>>>>>> Anatomy and Cell Biology / FEMR
> > >>>>>>>>>>>> McGill University
> > >>>>>>>>>>>> (514) 398-8084
> > >>>>>>>>>>>> mike.strauss at mcgill.ca
> > >>>>>>>>>>>>
> > >>>>>>>>>>>> --
> > >>>>>>>>>>>> To post to this group, send email to
> serialem at colorado.edu
> > >>>>>>>>>>>> To unsubscribe from this group, send email to
> serialem+unsubscribe at g.colorado.edu
> > >>>>>>>>>>>> If there is a different address on the following line,
> it is wrong.
> > >>>>>>>>>>>> ---
> > >>>>>>>>>>>> To unsubscribe from this group and stop receiving
> emails from it, send an email to serialem+unsubscribe at colorado.edu.
> > >>>>>>>>>>>> --
> > >>>>>>>>>>>> To post to this group, send email to
> serialem at colorado.edu
> > >>>>>>>>>>>> To unsubscribe from this group, send email to
> serialem+unsubscribe at g.colorado.edu
> > >>>>>>>>>>>> If there is a different address on the following line,
> it is wrong.
> > >>>>>>>>>>>
> > >>>>>>>>>>> --
> > >>>>>>>>>>> To post to this group, send email to
> serialem at colorado.edu
> > >>>>>>>>>>> To unsubscribe from this group, send email to
> serialem+unsubscribe at g.colorado.edu
> > >>>>>>>>>>> If there is a different address on the following line,
> it is wrong.
> > >>>>>>>>>>> ---
> > >>>>>>>>>>> To unsubscribe from this group and stop receiving emails
> from it, send an email to serialem+unsubscribe at colorado.edu.
> > >>>>>>>>>>
> > >>>>>>>>> --
> > >>>>>>>>> To post to this group, send email to serialem at colorado.edu
> > >>>>>>>>> To unsubscribe from this group, send email to
> serialem+unsubscribe at g.colorado.edu
> > >>>>>>>>> If there is a different address on the following line, it
> is wrong.
> > >>>>>>>>> ---
> > >>>>>>>>> To unsubscribe from this group and stop receiving emails
> from it, send an email to serialem+unsubscribe at colorado.edu.
> > >>>>>>
> > >>>>>
> > >>>>>
> > >>>>> --
> > >>>>> To post to this group, send email to serialem at colorado.edu
> > >>>>> To unsubscribe from this group, send email to
> serialem+unsubscribe at g.colorado.edu
> > >>>>> If there is a different address on the following line, it is
> wrong.
> > >>>>> ---
> > >>>>> To unsubscribe from this group and stop receiving emails from
> it, send an email to serialem+unsubscribe at colorado.edu.
> > >>>
> > >>>
> > >>> --
> > >>> To post to this group, send email to serialem at colorado.edu
> > >>> To unsubscribe from this group, send email to
> serialem+unsubscribe at g.colorado.edu
> > >>> If there is a different address on the following line, it is
> wrong.
> > >>> ---
> > >>> To unsubscribe from this group and stop receiving emails from
> it, send an email to serialem+unsubscribe at colorado.edu.
> > >
> >
> > <Faulty autogrid comparison.zip>
> >
> >
> > --
> > To post to this group, send email to serialem at colorado.edu
> > To unsubscribe from this group, send email to
> serialem+unsubscribe at g.colorado.edu
> > If there is a different address on the following line, it is wrong.
> > ---
> > To unsubscribe from this group and stop receiving emails from it,
> send an email to serialem+unsubscribe at colorado.edu.
> >
> > --
> > To post to this group, send email to serialem at colorado.edu
> > To unsubscribe from this group, send email to
> serialem+unsubscribe at g.colorado.edu
> > If there is a different address on the following line, it is wrong.
> > ---
> > To unsubscribe from this group and stop receiving emails from it,
> send an email to serialem+unsubscribe at colorado.edu.
> > <Faulty autogrid comparison.zip>
>
> --
> To post to this group, send email to serialem at colorado.edu
> To unsubscribe from this group, send email to
> serialem+unsubscribe at g.colorado.edu
> If there is a different address on the following line, it is wrong.
> ---
> To unsubscribe from this group and stop receiving emails from it, send
> an email to serialem+unsubscribe at colorado.edu.
>
>
>
> _______________________________________________
> 3dem mailing list
> 3dem at ncmir.ucsd.edu
> https://mail.ncmir.ucsd.edu/mailman/listinfo/3dem
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.ncmir.ucsd.edu/pipermail/3dem/attachments/20201008/4284ac3a/attachment-0001.html>
More information about the 3dem
mailing list