Monday, September 5, 2016

Is the Q Exactive HF less sensitive than other models?

Short answer:


At ASMS there was a rumor buzzing around. Earlier in the summer, 2 groups had found --individually -- on their assays, their QE Plus outperformed their QE HF in terms of sensitivity by limits of detection. Therefore, the rumor said, the QE HF wasn't as sensitive as the Plus.

I actively got involved in the first assay and resolved it myself -- it was a minor misconception on the parameters of the two instruments. The second was much more complicated and didn't seem to be my problem...until it was...and it irreparably destroyed a couple weeks of work I did in the spring.

The first one is the easiest to talk about (both thematically, and emotionally) but in the end, its the same issue.


This is a simple schematic of the QE to demonstrate how it works similarly to a triple quadrupole, but it'll do here as well. The focus here is the C-trap, which we control in the instrument software:

The importance of the C-trap in any hybrid Orbitrap system can not be understated. It is a critical and often misunderstood component of the system. In the newest instruments we have 2 ways of controlling this parameter in the instrument method software -- the AGC target and the maximum Injection Time (IT; which I often call "fill time").

The AGC Target is the goal. In this case, I am telling this QE Focus that the goal is for it to obtain 50,000 charges. That is 50,000 +1 ions; or 25,000 +2 ions; or whatever adds up to 50,000 total charges. That is the goal.

The Maximum IT is the backup plan. In this example, the QE is told to obtain 50,000 charges OR to fill for 57 milliseconds before it performs the HCD fragmentation and Orbitrap scan.

57ms is a ton of time to collect ions. Consider the fact that a modern QQQ (triple quadrupole) instrument running at maximum speed only spends 2ms on each target ion. Here, the QE can collect the same ion for >25 times longer than the QQQ.  {A QQQ has an advantage here because the ions actually physically strike the detector, while ions in the Orbitrap pass by the detector many many times, but that is a different conversation.} The fact is that in most experiments you won't ever need 57ms of fill time to collect 5e4 charges. You only need the maximum IT time on extremely low concentration ions.

Lets go to the Q Exactive Family Cycle Time Calculator! (Which you can download here).

QE Plus first:



In this case; MS1s are off, just MS/MS. Here I'm going at maximum speed. This is the most common settings for a QE; where the limiting factor is set to be the slowest part of the experiment -- the actual Orbitrap scan (which is 64ms on a QE or QE Plus); if you consider the 7ms of instrument overhead (to collect, fragment, cool, and inject) 57ms is the most efficient experiment. If you don't hit your AGC target, it will go to the next scan at 57ms anyway.  Lets call this the "High Speed" Experiment.

Lets take a look at the QE HF "High Speed" Experiment settings (please remember the Cycle time calculator is really an unofficial cycle time estimator -- its also a holiday here and I'm noticing this math isn't looking right, but I'm too lazy to check it):



Boom! We're flying here. We went from 12Hz to 18Hz or whatever, so the Orbitrap is faster!  We can get 20 MS/MS faster on the HF than on the Plus. If the AGC target is always easy to hit -- the HF is going to tear through 50% more scans than the QE Plus.

See the problem, though?? What if the AGC target ISN'T easy to hit? What if the maximum injection time is needed? Then the QE HF gets 57ms to collect ions; but the QE HF gets only 32ms --about one-half the amount of time that the Plus did. All the sudden, you're going to be looking at 1/2 the signal!!

The only way to ensure that the 2 instruments are running equally in terms of limits of detection or limits of quantification is to set the maximum injection time to the same number between the two instruments.

Imagine that you had all the same parameters -- same LC; column; sample concentration injection; AGC target; and Maximum IT -- how would that experiment turn out? The sensitivity will be just a little bit higher on the QE HF than on the Plus -- for an entirely different reason.

Imagine the top of your peak -- where the highest intensity of your target is coming out. This is the place where you are most likely to be hitting your AGC target.

Look at this peak I chose, literally at random, from the very first file I found in my Downloads folder. Also, please remember that if someone sends me a file to check out, 99% of the time something is wrong with it and they are asking me for advice on what might be the problem -- regardless, it is an okay example. I've labeled the scan numbers.

Check out lower on the peak --- the scans are further spaced than they are on the top. Near the bottom of this weak signal the Maximum IT is being used. Near the top of the peak, where we have the most ions, the AGC is being reached. The limiting factor at the top of the peak is the scan speed of the instrument.

Imagine now -- if I had a faster scanning instrument -- like a QE HF running this experiment -- here I would be getting 2x the number of scans at the top of this peak. Honestly, here I may have missed the actual top of this peak entirely. The max signal may have came out between 11739 and 11758 -- and I might have missed it. Even if 11739 was the highest that signal ever got, There is definitely some signal missed here between 11713 and 11739 that could have been picked up more accurately by a faster scanning instrument (or a less complex experiment).  Therefore, the faster scanning QE HF would get slightly higher sensitivity on the same experiment than a slower scanning instrument.

I believe this post deserves this image.

The QE HF isn't as sensitive as a QE Plus???


EDITS: Wow!  My 4th most read blog post ever? WTHeck? Okay. So just in case you think I'm making this stuff up. Please go to this paper, where the sensitivity of the QE HF is compared to the QE Classic. The HF can achieve the better fragmentation quality in ONE-HALF the time the Classic requires for the same sample and LC setup.


This study did not compare the Q Exactive Plus, but...this one did....



And...in this deep analysis, no deficiencies in the QE HF in terms of sensitivity were uncovered. Exactly the opposite.




6 comments:

  1. Hi Ben,
    I always thought that the AGC target was a number of ions....maybe the q-exactive is a particular case related to the CTCD electrometer. Do you know if the AGC target of the velos or fusion instruments corresponds also to a number of charges?
    Thanks :)

    ReplyDelete
    Replies
    1. I know for sure AGC target in all the QEs is #of charges. For the LTQ Orbitraps, I think (think) it is number of ions, not charges. Check out this in-depth study (J Proteome Res. 2013 Jul 5; 12(7): 3071–3086). They clearly state "ions" and I don't think they would mix up that distinction. Interestingly, the earlier Orbitraps and the Velos onward determine AGC in different ways, the logic behind it was detailed here (http://dx.doi.org/10.1016/j.jasms.2009.05.007) but it is detailed here (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2816009/) I'll have to ask someone about the Fusion, I think I know, but I'm not sure.

      Delete
  2. Thanks for the reply!
    An engineer from thermo told me that the target was a number of charges for all the instruments, even for the velos. For the velos, I also thought it was a number of ions still the multipliers counts ions and not charges.
    For the fusion i have read a patent, it seems to be a number of charges.
    The differences in AGC target between the different instruments are still unclear for me...if you have some informations, please let me know, I will be very happy !
    Thanks

    ReplyDelete
    Replies
    1. Definitely take the engineer's word for it. I asked a friend who knows a lot of mass spec physics and she also said it should all be charges, but asked me to check what kind of multipliers each device had to make sure. I did not do this. In terms of the AGC target numbers...I can guess exactly what you are asking...what I've been told (no reference for it, unfortunately) is that the ion transfer efficiency between where the ions are counted and the Orbitrap that the number of ions has to be decreased to prevent space charging. If I find a reference for that, I'll definitely put it up!

      Delete
    2. *words missing* that the ion transfer efficiency is much higher in the newer instruments.

      Delete
    3. One more person got back to me -- the LTQ Orbitraps are definitely counting charges from the prescan. I seriously pulled like 10 sources from the literature (good ones!) that say "ions" but if he said it's charges, it is charges. Maybe this should be it's own post...

      Delete