FAQ  •  Register  •  Login

Progressive increase in signal background

Forum rules
Please be polite and civil. We know that troubleshooting is vexing...
<<

AdeebR

Grand master

Posts: 169

Joined: Thu Mar 13, 2014 5:58 pm

Location: NYC

Post Tue Apr 30, 2019 4:14 pm

Progressive increase in signal background

Hi folks,

Over the past few months, I've had several people reach out to me with regards to an apparent data artefact where they've seen a progressive increase in signal background over the course of an acquisition (which we have also observed on sporadic samples). I wanted to share some results from some recent tests that suggest a possible cause: residual wash buffer contaminating the sample.

I've attached some slides summarizing some results and thoughts on this issue. Hope this is useful.

Adeeb
Attachments
190426_Wash_testing.pdf
(586.55 KiB) Downloaded 801 times
Adeeb Rahman
Icahn School of Medicine at Mount Sinai, NYC
<<

mleipold

Guru

Posts: 5795

Joined: Fri Nov 01, 2013 5:30 pm

Location: Stanford HIMC, CA, USA

Post Tue Apr 30, 2019 4:37 pm

Re: Progressive increase in signal background

Hi Adeeb,

Thanks for the detailed investigation and the PDF summary.

It's good to know that 30sec of water is typically sufficient to avoid the issue. I think that's about how long we typically go.

I advise my users to watch that it in a different way: the Wash solution has a Zirconium impurity which you can see in TOF view (masses 90/91/92/94/96; Mass 90 is ~50% of nat abund). Therefore, if you run water (or CAS) until those streaks go away, the assumption is that you've washed away the last of the Wash.


I wonder if some of the "occasional" issues we, you, and others have observed come down to the PSI not starting immediately when you press Sample Introduction "ON". As various PSIs have been replaced, I've seen an inconsistency (between PSI units) in the lag time between hitting "ON" and when the flow rate actually gets up to pressure and 30uL/min. One PSI might be 2-3 seconds, but I have one that's 5-10 seconds before it *actually* starts pumping (yes, it's really annoying, and so far I haven't found a way to adjust that, even in Service Mode).

So, if you hit "ON" and then start a timer for 30sec, but the PSI doesn't start properly pumping for 5-10sec, then your *actual* rinsing time would be 20-25sec, which might not be enough?


Mike
<<

dahern

Participant

Posts: 16

Joined: Tue Mar 31, 2015 3:45 pm

Post Tue Apr 30, 2019 5:35 pm

Re: Progressive increase in signal background

Hi Adeeb,

Thanks for this, it looks very similar to a problem we've been having with some samples.

Couple of questions if you dont' mind answering.

For the same staining panel, if the problem does arise does it always happen in the same channels for that panel? (ours does)
We routinely would run water for 5-10 minutes prior to starting acquisition but still see the problem - I'm guessing that you do too even when you run water for longer than 30 seconds?
Do you think this problem might be certain PSI's? Is it possible some wash solution remains in the fluidics system somewhere for some PSI (clutching at straws with this one I think).

Thanks!

David.
<<

EHaasDFCI

Contributor

Posts: 24

Joined: Thu Jun 15, 2017 5:34 pm

Post Wed May 01, 2019 2:08 pm

Re: Progressive increase in signal background

Hi Adeeb,

Thank you for sharing this. We have seen a similar occurrence in the past, but it is relatively rare for us. I was thinking along the lines (no pun intended) of your background being the ultimate cause of the signal drift issue. To what degree would you attribute this to extra wash solution in the lines or to inadequately washed samples prior to resuspension? As we all know, I'm sure, improperly washed samples tend to have higher background, which upon even just spinning down, aspirating, and resuspending the sample, the background is significantly reduced. Have you been able to pinpoint any background due to poor sample washing and link this to the issue presented?

Eric
<<

mleipold

Guru

Posts: 5795

Joined: Fri Nov 01, 2013 5:30 pm

Location: Stanford HIMC, CA, USA

Post Wed May 01, 2019 3:16 pm

Re: Progressive increase in signal background

Hi Eric,

At least in theory, background due to debris, poor washing, and mistitered reagents *should* be pretty stable across the time of the run. In other words, I wouldn't expect there to be the sort of gradual increase in signal intensity over time that Adeeb shows in his Aliquot 2 data: background should largely start out high and stay high, or start out low and stay low.

There can of course be minor variations in time (both up and down), or a decrease followed by spike in background right after a clog bursts through (though event rate would also go down and up in sync, which isn't shown in Adeeb's data). But nothing as consistent or as large a magnitude as Adeeb shows.


Mike
<<

OneJuan

Participant

Posts: 3

Joined: Wed May 01, 2019 6:31 pm

Post Wed May 01, 2019 7:00 pm

Re: Progressive increase in signal background

Adeeb, many thanks for investigating this. I had previously reached out to you and Mike.

In our case, we do not think it is wash solution as we see this increase in background in-between tubes of a barcoded sample, (without using wash solution between tubes, just CAS). Basically the background starts increasing as time progresses, then when we switch to a new tube, it starts from low background again, and increases, and so forth. We see it either if we stop acquisition, run CAS while the machine does the file processing, and load a fresh tube of the same sample, then concatenate files or interestingly, we also see it if we just stop sample introduction (without stopping acquisition), load a new tube and continue.
We mainly see it in 141Pr, albeit, since I know what to look for, you can barely see hints of it in other channels, but almost imperceptible.
I had previously reached out to Fluidigm, and they also thought that it may be wash solution, but the fact that we see it as we proceed with out run and we see it after switching to the new tube without washing in between.

I do not have permission to post attachments but I'll share the image once I can.
<<

AdeebR

Grand master

Posts: 169

Joined: Thu Mar 13, 2014 5:58 pm

Location: NYC

Post Thu May 02, 2019 12:26 am

Re: Progressive increase in signal background

Mike - thanks for the suggestion regarding the Zirconium monitoring; definitely something to add to our wash QC.

Eric - I concur with Mike. We have seen instances of elevated background due to overstaining, but this tends to manifest as constant high background signal rather than a progressive increase in signal over the course of the acquistion.

David - yes, it does seem seem like certain channels are more predominantly and consistently affected. We also do generally run water for longer than 30s, so I agree that when this happens during routine operation it's due to some sort of persistence of wash solution in the system.

My current theory (disclaimer: lots of hand waving to follow) is that this may somehow be related to the tubing that connects the Helios PSI to the grounding nut. This is actually two layers of tubing - the interior fused silica line that carries the actual liquid, and the 1/16" exterior sheath tubing that pretty much serves to increase the outer diameter to accommodate the grounding port nut. Ideally, the ends of these tubes should be completely flush, the end of the fused silica tubing should mate exactly with the port in the grounding nut, and all liquid should pass directly through from the silica tubing into the grounding nut. In practice, this often does not seem to be the case - the little compression fitting at the end of the line does not seem to be sufficient to the keep the tube layers of tubing together, and the inner tubing can sometimes slide back and forth. Under these conditions, the end of the fused silica tubing may no longer flush with port in the grounding nut, and so instead of liquid going through the port, it instead goes into the void between the two layers of tubing (we can see visible evidence that this is happening by observing liquid between the lines, which really should be dry). So potentially, what could be happening is that the ends of the tubing are not well aligned with the grounding nut port, wash solution enters the space between the two tubes and is consequently not effectively flushed out when running water and therefore persists at levels sufficient to contaminate the sample.
Adeeb Rahman
Icahn School of Medicine at Mount Sinai, NYC
<<

DMcDonald

Contributor

Posts: 21

Joined: Thu Nov 26, 2015 10:14 am

Post Thu May 02, 2019 2:49 pm

Re: Progressive increase in signal background

Hi Adeeb,

Thank you for this, it's come at a time troubleshooting very similar data. One thing I noticed and was EQ bead channels were affected to a greater extent. Presumably the EQ beads are also being stripped of some of the metal but it's less obvious visualising these in the higher log decades. This may help explain why the affected channels are spread across the board, perhaps too if your hunch about the sample line tubing is correct then some tuning solution could also be hanging around. Unfortunately a lot of our new sample lines come at the wrong length and leak backwards from the grounding nut, I trim them flush with a small guillotine once connected to the PSI which seems to help a lot. Another source could simply be the residual wash on the side of the tubing dipping into the samples, if a large volume tube of wash sol is on followed by a smaller sample then a decent proportion could drip down into the sample with time. We see something similar on fluorescent analysers where large volumes of bleach have been used.

Again thanks for the really helpful data.

David
<<

mleipold

Guru

Posts: 5795

Joined: Fri Nov 01, 2013 5:30 pm

Location: Stanford HIMC, CA, USA

Post Thu May 02, 2019 3:08 pm

Re: Progressive increase in signal background

Hi David,

Could you share some of your affected EQ bead data, vs some unaffected EQ bead data (ideally from the same day)?

To confirm: when you say "EQ bead channels were affected to a greater extent", do you mean those channels on the actual EQ beads themselves, or do you instead mean the same mass channels when used for antibody/probes on your samples?


The reason I ask: from lab meetings way back in my Toronto postdoc days, I don't recall a ton of leakage of Ln *out* of polystyrene beads, even under acidic conditions. For example, see Fig 5 here (timescale weeks): http://dx.doi.org/10.1021/ja9052009 In another paper ( http://dx.doi.org/10.1039/b921770c ), they did get lanthanides quantitatively out of beads, but did so by drying the beads, grinding them, then using concentrated nitric and microwave digestion.


Has anyone tried a digestion of the EQ beads in 2% nitric or Wash Solution, then run in Liquid Mode? I think you'd only have to do this for 30min or so, since at least some of our samples would be on this timescale.


Mike
<<

OneJuan

Participant

Posts: 3

Joined: Wed May 01, 2019 6:31 pm

Post Mon May 06, 2019 4:14 pm

Re: Progressive increase in signal background

This is a screen cap of a post that I submitted Thursday, for some reason it didn't go through. The time scale, the first 2 segments of acquisition are about 2.25 hrs, the final segment is around 1.9 hrs. In this case there is no concatenation, just sample introduction was stopped to change tubes. It looks the same (after normalization/concatenation) if acquisition is stopped, CAS is run while the computer processes the file and then a new tube of the same barcoded sample is set to acquire.

Image
Attachments
Screen Shot 2019-05-02 at 7.49.20 AM.png
Next

Return to CyTOF troubleshooting

Who is online

Users browsing this forum: No registered users and 10 guests