Questions about Tuning report

[zhangyiwen]

Hi, I am just start to learn CyTOF, and I have three questions in tuning instruments.

1.Why oxide check use Gd155/Tb159 ratio, there is no M+16 relationship between them. Why not use Ce140 to CeO 156（or Ce140 with Gd156）or Nd146 with Dy162？
2.What is “Dual slope” in Dual calibration analytes？(especially Dual means what？) And is “R2” stands for correlation coefficient？Ir has a R2 of 0.26044，not very good for a Correction factor in my opinion.
3. In “QC analytes” ，what are the relationships between “mean intensity” and “mean duals” or “mean pulses”？ And what does “Dual counts” mean？

These are the questions， maybe very basic for you, but they bother me for a long time, hoping to get the answers.

Best，
YiWen

[mleipold]

Hi Yiwen

1. In general, the CyTOF should be tuned in the factory so that 159Tb and 169Tm are the highest sensitivity. 139La is the most easily oxidized lanthanide, so if you control that oxide below 3% (139La + 16O = "155Gd"), you're automatically controlling all the other oxides below 3%. In other words, the Tuning protocol is set up to maximize true signal in 159/169, while keeping the "155Gd" oxide at an acceptable level.

2 and 3: I'll leave someone else to discuss R2. However, the Intensity/Pulses/Dual (counts and slope) are discussed in the manuals, and also in a review we wrote a few years ago: https://doi.org/10.1002/cyto.a.23621 (see Figure 2 and "Data Acquisition" section)
- in short, for accurate measurement, Pulses are counted at very low signal, and Intensity is counted at higher signal. Rather than have two different units of measurement depending on the magniture of your signal, Dual is a correlation between Pulses and Intensity to unify the units across the entire scale.


Mike

[CRStevens]

Hi Yiwen,

If you got into CyTOF early they used to give you a great hard cover training manual that got really in depth on many of these topics. Unfortunately I dont know of a good online resource for them. I'll try to help best I can. What I can do, though is when I get back in the office I could make a PDF from my old docs and post it here for newcomers.

#2
When you run the tuning solution if plots intensity of 159Tb against dual counts of 159tb. The slope of that regression line is your dual slope. Way back, you used to be able to manually tune your instrument that allowed for higher intensity but lower counts, or vice versa. This was helpful if you wanted say higher intensity to detect weak signals. They have locked this down for an optimal slope of 0.3. There is a +/- to this and it should be close to that 0.3 to be properly tuned.

#3
more complicated, but its the conversion of the analog to digital data format and converting it to based on the pulses, intensities, and counts. see attached from CyTOF 2 manual
Really the mean duals is all you have to pay attention to, but will upload the explanation from the old training manual when I get a chance for you.

[zhangyiwen]

Thanks Mike and Chad for your answers,

@Mike
On the basis of your reply, can it be considered that there is no actual "Gd" element in the tuning solution, so that the intensity of Gd is three orders of magnitude lower than other elements(Except Ba, because "the instrument is most sensitive
to masses in the AW 159–170 range; lanthanide masses outside this range may have up to fourfold lower signal intensity" according to https://doi.org/10.1002/cyto.a.23621)？

To my understanding,
1.one can also identify "Oxide ratio" as LaO(155) to La(139) to set a more stringent standard, and
2.because the instrument is not like normal ICP-MS which has the specialized window (formula) to measure oxide and double charge, so that use other element name ( e.g. Gd) to replace LaO(155) ?
Is this right? But if TOF get correct mass of LaO (m/z = 154.905753 u, oxygen element is calculated according to natural distribution) and Gd (mass of atom = 154.922619 u), there are still differences of 0.016866 (154.922619-154.905953).


@Chad
I understand pulses, intensities, and counts much better. It will be very helpful having doc/PDF reading materials, please attach them in your convenience.

Yiwen

[mleipold]

Hi Yiwen,

Correct. There is no actual Gd155 in the Tuning solution: the 155 mass is LaO.

Yes, formally it would be more stringent to define oxide ratio as LaO/La (155/139). I have no idea why DVS/Fluidigm/SBio decided to do it that way.

The "name" of mass 155 is irrelevant. By happenstance, there is no other element with an isotope at mass 155 (like now Tb is the only element with a mass 159 isotope). Therefore, by Fluidigm's software scheme, you have to select "155Gd" in order to monitor/measure the 155 mass channel. But on a fundamental level, that is a software labeling issue, not a hardware issue. If you want to measure the 113 channel, you could choose 113In or 113Cd, and the TOF+detector wouldn't "care". I personally feel it's misleading to your future self and anyone else who uses your data to "mislabel" the element label, as the spillovers of 113In are not the same as 113Cd. But, in terms of data acquisition, it's irrelevant.

Formally, yes, the correct masses of true 155Gd vs "155Gd" (139La+16O) are slightly different. I'll let someone from Fluidigm or some other hard-core MS expert make a deep-dive formal response to the differences between standard ICP-MS measurements vs those from a CyTOF, at the level of 0.02 mass units.

That said: the "mass" is calibrated during the Mass Calibration step of the Tuning process. This is based off the 133Cs and the 193Ir masses in the Tuning, which are used to define what mass 133 and mass 193 are (ie, the TOF arrival times of ions of those masses), and then the mass/arrival times of all the other masses are calculated off that framework. So, when we're "measuring" mass 155, what we're actually measuring are the ions that arrive at the TOF arrival time calculated to correspond to mass 155. This arrival time can drift over long runtimes due to some things like vacuum levels in the TOF chamber: as such, this is one reason why long runtimes usually need a redo of the Tuning. This was particularly an issue for CyTOF1s which in my hands never really completely stopped drifting; in my hands, Helios (CyTOF3) are much more stable, and so retuning is usually more to do with DV changes than Mass Cal changes.

So, given an arrival time of X for mass 155, the software then captures the signal for a TOF window +/- y: so, from X-y to X+y. The sum of that signal (roughly area under the curve) is what is then reported as the signal for mass 155. My *guess* is that the 0.02 mass unit variance between true 155Gd and "155" (139La+O16) would amount to a shoulder on that ion peak, but still fall in the X-y to X+y calculation window to be summed as mass 155, and still retain baseline resolution between adjacent masses (between 154 and 155, and between 155 and 156). Lower masses like 133Cs have narrower TOF arrival distributions (ie, the TOF peak is narrower) than higher masses like 193Ir (see attached PDF), and yet there's still baseline resolution between mass 193 and mass 194.


Mike