Quan on the QE Part 4: Data independent acquisition

Haven't read the rest of it?   Here is part 1

This is part 4 of my monologue on (what I was calling targeted) quan strategies for the Q Exactive.  Last week someone told me that he never says the full name of this technique to collaborators outside of the field, as biologists used to hypothesis driven research seem appalled that we would do an experiment that was not dependent on data, lol!

There are two types of DIA and they need to be addressed separately, because they are more than a little different.  We have DIA and multiplexed DIA, or MSX-DIA.

DIA is a strategy that is extremely similar to a technique on other platforms known at SWOTH, or something.  In this, we collect the fragment ions from within a relatively large mass window, and synchronously move to the next larger, non overlapping mass window.

For example, we may start by fragmenting all of the ions from 400-425, the next scan will fragment all of the ions from 425-450, and we'll typically go up all the way to 800 or 1000.

Advantages:  Extremely easy to set up on the instrument side of the Q Exactive. Gives you access to an almost limitless number of targets with increased sensitivity over that of a full MS1.  The data can always be re-mined when other targets of interest are identified.

Disadvantages:  All of the work needs to occur on the processing end.  Lower dynamic range than other methods here, such as T-SIM.  Specificity may be an issue in complex samples.  On large numbers of windows, cycle time can become an issue.

There are definitely places for DIA and I've recommended it to some groups recently.  For example:  if you are in a lab where you have an ultra sensitive discovery instrument like an Orbi Elite searching for biomarkers and you are validating them on multiple patient samples, DIA may be a great place to be when that Elite pops up an additional biomarker.  Rather than re-running everything, you simply pull out these data files and drop them into Pinpoint or Skyline and search for this new marker.  No reruns.  Better dynamic range than MS1.

As is SWOTH, this technique is not really compatible with UPLC.  If you have a 6-second peak, you aren't going to be able to make it through a full set of targets from 400-800, though the 64ms cycle time on the Q-Exactive is sure a lot faster than the mandatory 100ms dwell time the "super fast" SWOTHY instruments don't tell you about in their marketing literature, lol!

Specificity can be a problem here.  In DdA (data dependent), we go for larger peptides because of the low likelihood of their existence in random events or other organisms, right?  The chances of 2 amino showing up side-by-side is 26^2 (or 1 in 676).  The chances of 8 amino acids showing up in the correct order is 26^8, or 1 in 206 billion (yes, I'm ignoring reverse sequences, but the increased incidence in stop codon sequences gives me justification while this espresso kicks in.  And yes, science conscience, I'm also ignoring sequence homology effects from conserved evolutionary events, but this number is still pretty freaking high).

Now, if we our only data is "this ion has an m/z of somewhere between 400-425 and we see fragment ions 1,2 and 3 that support its existence" we are relying on a good bit less data than we get from a traditional DdA where we know the m/z within a 10 ppm window and we have fragments from the whole sequence.  In most cases, say ions 1,2, and 3 are all 3 amino acids long ((26^3)x3 (or, probably,^3?)) gives us some high level confidence of specificity, but we'd really need multiple fragment ions resolved well.  I sure would trust it a whole lot more if I had high res accurate mass on the QE rather than on old SWOTHY, but that is probably another talk for another time.

Okay, I hate to break this up yet again, but I got kind of long winded here.  There will be a Part 5, where we crank up the specificity and the speed with MSX-DIA.  I might even do it today since I won't be going anywhere until a snowplow comes through...