My Opera is down more that up

[Michelle Ocana]

Hi All,
I think I will kick off the discussions.
The problems I have encountered have been many but the outstanding
issues are with skewanalysis at 60x and cam3. Opera fails skew but
doesn't throw an exception at 60x. I find that the skew was off only
after acquisition. I'm collecting hoechst staining in both exposures
now as a control for skew and it's pretty bad. It can be off
considerably for the particle I'm resolving. I'm looking at pixel
shifts of 5-25 pixels. The most frustrating is the skew is different
for every field. Something is moving, or at least not returning to the
same position for each exposure.

I'm also collecting in Cam3 which doesn't appear collimated. The beads
look out of focus and larger than in the other cameras. I had the
service tech bring the cam3 up and down in z and it doesn't fix the
focus issue leading me to believe this is a collimation problem.
Evotec says they know about this and it is corrected in skew analysis.
It's not corrected. I feel like I should say that again, it's not
corrected. I can not use Cam 3 to measure anything (Really how bad can
it be. It's out of focus so pixel information is incorrect and the
image is larger than it should be so the spacial information is lost.
I can't measure size or intensity). I use it for marker only (to say
yup, there is stain there). If I can't use cam3 to measure I was going
to ask them to remove it and put the DM in the DM position with the
510 and 580nm and place the CY5 BF filter in cam2 position. They can
take Cam3 home with them. I will lose a camera but it seems a moot
point since it really doesn't work anyway.

After many days of webx, onsite service, and many many skewanalysis,
Evotec said that they are looking into z position as the problem with
skew. In other words, their pifoc isn't so good and it may be
affecting the skewanalysis. So they adjusted the AF parameters and the
skew seemed better when I acquired 3 fields in 1 well as a test. I
think the changing of the AF was helpful for skew, they must have
tightened the tolerance for focus. It seems like the AF needs to be
really precise at 60x. Too precise for their pifoc program. The plate
will fail to focus on more that 30% of fields in a plate. Guess I will
be seeing a lot more of my engineer. I think all of this could be
overcome if they changed their skewanalysis to run more than once at 1
z position before saving.


The questions I want to post to the community are;
How often do you use 60x in plate acquisition?
Does everyone have trouble with skewanalysis at 60x when trying to
resolve small (1-2um) particles?
Is skewanalysis failure to overlay seemingly random per field for you
at 60x?
Does skewanalysis correctly adjust the size of the bead in cam3 at
60x?
How accurate would you say your pifoc is at 60x? is there a tolerance
of .1um, .5um, more?
Does it often fail to AF?

[gbo...@gnf.org]

Michelle,

Thanks for posting the first discussion. First could you indicate
whether your Opera is a TEHS or a QEHS? From your description of the
problem, it seems that camera 3 is the non-confocal channel? If so, I
think the “collimation problem” is indeed typical of the system
(something to address in future version of the system). If your
channel 3 is confocal, there is typically a little play between the
channels, but it is never as drastic as between the non-confocal
channel and the confocal ones.

Could you possibly post some pictures of the beads (before and after
the skewcrop correction) so that we can see what your problem is?
Unfortunately, we only have a 60x on our TEHS, and a 40x on our QEHS.
I will try to get some images next week before and after skew on our
two systems at the highest resolution (60x/40x binning of 1).

I can tell you that from experience, I never had any problems at 40X
on our QEHS, and I think that things worked OK on our TEHS at 60x. For
the focusing, you also have to take into consideration the tolerance
of the plate you are using. I believe that you can define different
tolerance in the lense.ini for your different objectives and have a
plate/objective definition as well. This might help fine tune the
focusing for a given plate type at 60x.

In theory the auto-focusing should be better at higher magnification,
because the pifoc pics are crisper. If this does not look to be the
case, perhaps your pifoc is dead, or the focusing mechanism in your
Opera is bad (need to replace the microscope base). I would also
recommend to have the P&E technician level the stage (make sure the
right side of the picture is at the same confocal height as the left
side).

When I have observed the problem you describe (ie. Successful skew
crop that does not align pictures correctly afterwards), this problem
was consistent throughout every well. For the 1-2µm beads, let me do
what I can. I think we typically use 5µm beads to do our skewcrop at
40x and 60x. I will investigate how good the repositioning is by
collecting multiple stacks of beads at the same location multiple
times.

At 40x, on our QEHS, the autofocus never fails. I can check what
fraction of the plate fails at 60x on our TEHS.

Finally, I agree with you that the skew/crop analysis should be done
on more than one field and perhaps also on more than one z-plane. I
have sent a long email about this to Evotec at the time (over 2 years
ago). Furthermore, this entire skewcrop/ref image correction should be
acquired in a fully automated fashion, and should be validated
(measure how good the correction is). I hope that other customers will
request a similar feature.

[gbo...@gnf.org]

Michelle,

I posted 2 set of Opera images before and after skew analysis using
the 60X on our Opera TEHS (with a binning of 2). It seems that there
is no problem here. I imaged a full 384-well plate (excepted for the
well on the edges) and the overlay seems to work correctly in every
well.

That being said, I was not able to generate a skewcrop analysis with a
binning of 1 for the CCD (this is probably linked to the software
update to Evoshell 2.0, and should be fixed soon).

In addition, even though there were no autofocus failure, many well
were out of focus (probably half of the well). I suspect again that
something needs to be tuned with the autofocusing mechanism for the
60x. Keep you posted when this is fixed.

For the skewcrop analysis I said I used 5µm beads at 60x, I meant
2.5µm beads.

I hope these images can help you compare what you get with your
system. A QEHS should work much better!

Finally, I am sure that you have checked all of this, but on the TEHS,
the drop of water is not always correctly formed. This might cause a
problem with your system, so you may want to check that the water
dispensing is working properly. Also, double check that no air bubbles
is trapped in the system by priming the line for ~5" prior to imaging.

[tonyc]

Hi Michelle,

This sounds as if it's been quite annoying. We've beeen using the 60x
with the cellcarrier plates for a while with pretty much no problem.
One issue I had was that the objective hadn't been seated properly. It
refused to do a skewcrop until it was reseated. It did focus OK though
and you couldn't tell that the objective was not seated from the
image. It gave a "Ch0>Ch1" type error when the skewcrop failed.

Also, what plates are you using? Could it be the plate definition is a
little off and the usual variations in plate manufaturing are taking
it outside the range that the Opera's expecting?

Hope this help.

Regards,

Tony


Tony J. Collins, Ph.D.
McMaster Biophotonics Facility
Dept. Biochemistry and Biomedical Sciences HSC 4H21A
McMaster University, Hamilton, ON, L8N 3Z5
tcol...@macbiophotonics.ca www.macbiophotonics.ca