FAQ  •  Register  •  Login

Helios injector on CyTOF2

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 Fri Jan 20, 2017 6:48 pm

Helios injector on CyTOF2

Hi all,

We recently did a little testing to see whether we could stick a Helios injector on our CyTOF2 to gain some of the advertised benefits of smaller ion cloud size, which should provide higher signal intensity and reduced doublet incidence.

I set our flow rate to 30uL/min and successfully optimized our makeup gas at this level to get a nice boost in signal intensity with tuning and with beads. The bead CV was also terrific. However, when running cell samples we strange saw a reduction in cell signal intensity for most markers, despite the fact that the bead signal intensity was higher. I've attached some example plots:

Injector testing.tif


Does anyone have any thoughts as to what might be going on?

I suspect this relates to the minimum event length cutoff, which was set to 10. As expected, the Helios injector produced events with a much shorter event length, and it almost seems like they might be "clipped" off. This does raise an interesting question regarding what exactly the minimum event length threshold is doing. I was under the (perhaps erroneous) impression that the event length threshold should act like the signal threshold on a cytometer, where if an event signal doesn't exceed the threshold that event doesn't get detected. However, if an event does cross the minimum threshold then all the signal associated with that event is collected. This would mean that an over-agressive threshold would reduce the overall number of detected events, but should affect the event-associated signal intensity; however in this case it seems like there may in fact be an effect on signal intensity as well.
Adeeb Rahman
Icahn School of Medicine at Mount Sinai, NYC
<<

mleipold

Guru

Posts: 5792

Joined: Fri Nov 01, 2013 5:30 pm

Location: Stanford HIMC, CA, USA

Post Fri Jan 20, 2017 7:01 pm

Re: Helios injector on CyTOF2

HI Adeeb,

In my experience, yes, hitting the EL limit affects *number* of events, but not *signal* of events. I've had a couple customers who do their own staining, and messed up the Ir step to make it too dim. On the Helios, this caused their EL to average about 7; with a lower limit of 10, this caused a number of their events to not get written to the FCS file. As you know, you can always reprocess the IMD with a lower EL to generate a new FCS file; in this case, they got about 15% more cell events (the number's a bit "off" because they also got proprortionally more debris as well). We had to set EL-lower =5 for all future experiments to ensure they would get their events even if they screwed up the staining again.

To confirm: for this comparison, you stained cells, then split them into two aliquots? You ran one aliquot with the v2 injector, then shut down, swapped out for the Helios injector, then retuned, *then* ran the second aliquot with the Helios injector?

If you reprocess your IMD with EL=5, does your signal move back? I would not expect that it would, based on the EL argument. But I don't have an explanation for your data....


Mike
<<

AdeebR

Grand master

Posts: 169

Joined: Thu Mar 13, 2014 5:58 pm

Location: NYC

Post Fri Jan 20, 2017 11:52 pm

Re: Helios injector on CyTOF2

Hi Mike,

So it sounds like your experience is consistent with my understanding of how things are supposed to work, which shouldn't account for the lowered cell signal intensity. Unfortunately we didn't save the IMD file for this test so we can't reprocess the FCS.

And yes, your summary of the experimental design is correct: it was a single stained sample, divided into two equal aliquots, one of which was run on each of the injectors following tuning. The aliquots were stored in CSM and only re-suspended in H20 right before acquisition, and we also tested another samples where we switched around the order in which we tested the two injectors and got a similar result.

Adeeb
Adeeb Rahman
Icahn School of Medicine at Mount Sinai, NYC
<<

noelcasey

Participant

Posts: 8

Joined: Fri Apr 08, 2016 8:00 pm

Post Sun Jan 22, 2017 1:13 am

Re: Helios injector on CyTOF2

Adeeb,

Do you have example numbers/settings for your tuning of the cytof2 with both injectors. The narrow injector leads to an increase in argon gas velocity and a reduction of droplet/cell residence time within the plasma. So the optimum settings for tuning the icp can be different when compared to a large injector diameter. What is the size of the cells used in this test? This could in part be related to more cells not being completely vaporized in the plasma when compared to the much smaller beads.

Best,

Noel
<<

GregBehbehani

Master

Posts: 85

Joined: Tue Apr 12, 2016 10:17 pm

Location: The Ohio State University, Columbus, Ohio

Post Mon Jan 23, 2017 3:21 pm

Re: Helios injector on CyTOF2

Hi Adeeb,

Kudos to you for trying this and telling us about it. I don't know that I can give you much help except to suggest that you check all of your bead oxide parameters, as this will be your best indication of the plasma temperature and ionization.

I agree with Mike, the minium EL setting doesn't seem to change the overall signal in our hands; it just affects which small events get counted. We routinely run at a minimum length of 7 or 8 on all samples on our Helios, but back when we were playing around with it, we wouldn't see any change in the signal of events with an EL above the minimum regardless of the setting. Do keep in mind that the minimum event length and lower convolution threshold are related. The lower you set the lower convolution threshold the earlier the events will start and the longer the event length will be, I could imagine some combinations of lower convolution threshold and minimum EL that could create some of these effects, but it's hard to believe that you could get effects this large.

The thing that could account for some of your findings would be background signal. If some of your channels have background signal in solution, then this signal will get added to the integrated event signal in direct proportion to the EL. Theoretically this shouldn't happen if you're using background substraction, but I'm not sure how accurate the background subtraction really is.

A few other thoughts:
1. It looks like your DNA signal is higher with the Helios injector. Any idea why?
2. The other parameter you might want to pay extra attention to is the current. I notice that the Helios requires a much higher current setting than the CyTOF 1&2 machines I used at Stanford. I assume is due to the fact that the ions are moving faster (as Noel pointed out) and need more of a electrostatic "push" to get them focused in time. Theoretically, you might also need a different "turn" voltage for the same reason, but I can't think of any reason the ions would behave differently between cells and beads once they get into the machine.
3. Finally, I would just double check the exact setting you had after you re-tuned the machine to be sure nothing else (like the detector voltage) ended up changing during the tuning.

Good luck with this, I'm excited to hear what you figure out.

best,

Greg
<<

AdeebR

Grand master

Posts: 169

Joined: Thu Mar 13, 2014 5:58 pm

Location: NYC

Post Wed Jan 25, 2017 11:44 pm

Re: Helios injector on CyTOF2

Thanks for your thoughts, Noel and Greg.

You're right that the optimal gas settings for the new injector were quite different:

The CyTOF2 injector autotuned at Neb gas = 0.17, Makeup = 0.83, Current 4.0, DV -1951, which gave a Tb dual of ~776,000
The Helios injector autotuned at Neb gas = 0.17, Makeup = 0.48, Current 5.0, DV -1960, which gave a Tb dual of ~960,000

How do these values compare to what folks are getting on the Helios?

The cell samples we used for testing were PBMCs and whole blood.

I have a hard time understanding why ion clouds generated from cells vs beads would behave differently, which makes me think that this somehow relates to something that's happening pre/during ionization (or some sort of data processing artifact). I find Noel's suggestion of incomplete vaporization of the cells interesting. I wonder if cells have an intrinsically lower ionization efficiency than beads? I also noticed Greg's observation of the ~30% higher DNA signal with the Helios nebulizer and am not sure how to explain this since every other parameter except for DNA was lower.
Adeeb Rahman
Icahn School of Medicine at Mount Sinai, NYC
<<

mleipold

Guru

Posts: 5792

Joined: Fri Nov 01, 2013 5:30 pm

Location: Stanford HIMC, CA, USA

Post Thu Jan 26, 2017 3:44 pm

Re: Helios injector on CyTOF2

Hi Adeeb,

Your Helios values are basically what I'm getting on my two Helios.

If you decide to leave the Helios injector in your v2, keep an eye on your Current setting: I have seen a much greater "burn-in" for the Current with a new skimmer-reducer than on v1. By that, I mean a brand-new skimmer reducer may have a value of I=5, but within 2 weeks, it will climb to I=7 or so, and never get below that again (ie, it's the "new normal").

The HIMC has a v2-to-Helios, but since I don't use that particular instrument that often, I don't know how it was affected. However, talking with a couple other people who got v2-to-Helios upgrades, they saw a similar ~2 Current unit increase after their upgrades. I'm not sure why it's happening, or why Helios machines in general seem to need a higher Current than v1 or v2. The best explanation I've heard is that if the ion cloud is more compact (same number of ions, just smaller diameter), maybe the skimmer-reducer needs more Current to apply the attraction voltage to efficiently get the ions into the skimmer-reducer orifice.

But that's a guess from a couple people....


Mike
<<

noelcasey

Participant

Posts: 8

Joined: Fri Apr 08, 2016 8:00 pm

Post Sat Jan 28, 2017 9:43 pm

Re: Helios injector on CyTOF2

Hi Adeeb,

Thinking of time and distance is the key to understanding how the ICP optimizes. Based on an argon gas velocity of 20m/s, it takes 2ms for a droplet,cell, bead to pass through the ICP. So droplets, beads, cells would have to complete the process of desolvation, vaporization, atomization and ionization within that time frame (see attached image). If different sized droplets/cells/beads take a longer or shorter time to complete the process they would optimize at slightly different operating conditions. Typically a droplet of less than 10 microns is efficiently ionized within the plasma.

So choosing a smaller inner diameter injector increases the gas velocity and reduces the residence time within the plasma. This requires a lower gas flow rate to obtain similar results when compared to a larger inner diameter injector. You could test this if you measured different sized cells labelled with the same element or beads with cells compared to beads alone.

A couple of good references about the origin of signals within the ICP are below.

Investigating the Fate of Individual Sample Droplets in Inductively Coupled Plasmas
John W. Olesik, Applied Spectroscopy, Vol 51, Issue 5, pp. 158A - 175A
http://journals.sagepub.com/doi/abs/10. ... 2971940792

Behavior of Bacteria in the Inductively Coupled Plasma:  Atomization and Production of Atomic Ions for Mass Spectrometry, Fumin Li, Daniel W. Armstrong , and R. S. Houk, Anal. Chem., 2005, 77 (5), pp 1407–1413
DOI: 10.1021/ac049188l
http://pubs.acs.org/doi/abs/10.1021/ac049188l
Attachments
Sequence of Processes in the Inductively Coupled Plasma that lead to ICP signals.jpg

Return to CyTOF troubleshooting

Who is online

Users browsing this forum: No registered users and 9 guests