Improving Challenging Separations With SLIM-Based Ion Mobility Mass Spectrometry

SLIM was invented at the Pacific Northwest National Laboratory, and MOBILion is the exclusive licensee to the technology for around nine years now.

What’s really cool is that, because it’s electronically driven and uses printed circuit boards, we can change the electrode pattern and the electronics to make it do different things. At the 10,000-foot level, it makes liquid chromatography–mass spectrometry (LC-MS) workflows better; if you insert it into an LC-MS workflow, it will give you a better, faster, more reproducible, more sensitive, higher-throughput analytical answer. In some cases, we say it can reveal what others leave unseen—it sees things that traditional LC-MS cannot. The form factor is customizable and tunable, so it's applicable across all LC-MS workflows and all flavors of mass spectrometry.

Over the course of the past nine years since we've licensed the technology, we've developed products and shown proof of concepts. The first was our MOBIE® platform—our commercial product in the high-resolution non-proteomics segment of the market. Last year, we announced a new operation mode using the same core technology, just a different configuration of electrodes and different electronics, to make it applicable for the discovery proteomics market in the high-resolution space. This year, we're giving people a sneak peek at our low-resolution integration, which uses SLIM in front of triple quadrupole instruments to give higher throughput and higher sensitivity—around 100x more sensitivity than conventional triple quad workflows—so it's elevating the LC-MS space holistically.

In mass spec, we ionize samples, so you are always looking at ions that represent the analytes in your sample. For better sensitivity, you want all of these ions to be detected; if you’re losing ions, then you’re not detecting things that were in your sample.

If you look at traditional LC-MS workflows, they're “lossy” in terms of the ion utilization. That is driven by quadrupole function; whether you're using a Q-TOF high-resolution MS or a triple quad MS, that quadrupole filters and scans—it's not a continuous ion utilization component. We offset the deficiency of the quadrupole by being able to preserve all of the ions, all of the analytes, and maintain essentially 100% ion utilization, so you don't lose anything. 

It’s a lot to do with the analytes, so you see this in the PFAS/environmental space, the food space, in the pharmaceutical space. Really, it’s across the board. LC separates in a fundamentally different way—it is chemical separation, whereas SLIM is separation based on a physical property of a molecule. Just by adding our type of separation to the LC type of separation, you get the ultimate separation power.

Genentech presented at our ASMS breakfast seminar earlier. They used conventional LC-MS for whatever they were looking at, and they saw one peak. Then they analyzed that sample with our instrument. Across the three different examples they gave, they saw 7 peaks, 9 peaks, and 10 peaks under that 1 peak.

It is taking characterization to the next level—where with LC-MS you think you've got 1 analyte there, you add in our separation technology, and you realise that there are actually 10 analytes there instead. Whether these are PFAS, sugars or food molecules, pharmaceutical molecules, this difference is mindblowing. Think about this error that exists, that you didn’t even know was there. If these 10 peaks all represent different molecules, in the case of a drug, those are some molecules that maybe have good efficacy or maybe are contributing to some negative adverse side effects. You really do have to know the chemistry of that drug at this level, and that is what we tap into.

Proteomics is very diverse; there are a lot of people trying to do a lot of different things. For PAMAF—our parallel accumulation mobility aligned fragmentation mode—it’s especially targeted towards high sensitivity and high throughput workflows.

In this mode of operation, you get around a 10x increase in sensitivity, which really works well for single-cell analyses or sample-limited analyses, where you don't have much material and you need to get as much as you possibly can out of it. That is one of the advantages of that PAMAF mode of operation—being able to make the most of everything that's actually coming into the system.

There's value in that PAMAF mode of operation for these specific situations where sensitivity is so critical. But in proteomics in general, by and large, people are throwing away a lot of information. In a typical protein expression experiment, you might get up to 6, maybe 10 peptides, and conclude that you know what that protein is. But if you sit and listen to a lot of the conversations and talks that are happening here at ASMS, everybody's talking about post-translational modifications. It's not just a case of asking “is the protein there, and how much is there,” it's moving towards also asking “what has happened post-translationally to that protein?”

Our technology, because we get more precursors, because we get more ions that are going through, that translates to more peptides, which translates into more coverage of each one of those proteins. Fundamentally, you know more about the proteins that you've seen during your analysis. Not only does that help you answer the initial question you had, but the translation gets much better. That's where there have been a lot of problems in proteomics in the past—you make a discovery but then it doesn't go anywhere; another discovery and then it doesn't go anywhere. A lot of that has to do with the fact that you didn't fully understand the protein in the first place. With something like the SLIM technology, you get considerably more information about all the proteins you find, which gives you a better head start when you move into that translational next step.

It has been difficult. Bear in mind that we’ve been at this for nine years and have been working with prototypes and products for five of those years. But I always say, it is a blessing and a curse. It is a blessing to be able to elevate mass spectrometry workflows across all of these diverse applications and across all different types of mass spectrometry instruments. But is it also a curse, because where do you begin?

The product and the configuration that we have for non-proteomics applications is different to the operation mode and the printed circuit board electrode design that we have for proteomics. We have to make bespoke products that function absolutely in sync with the operation mode of that particular flavor of mass spectrometer for that specific application.

The good thing is, a lot of the hardware can stay the same; it is not like we’re starting from scratch all the time. But there are real nuances to different workflows, so it just takes time. Truly, there is an endless list of opportunities and applications, and we can’t do them all at once, so we have had to pick and choose what comes first, second, and third. But also, it is mindblowing what our install base is already doing with the MOBIE product—they are continuing to explore, make discoveries, and add new applications to the list all the time.

For me, there are a couple of very exciting things that I see in the future.

Firstly, because the workhorse separation part of our workflow is electronically driven and analyte agnostic, it can be more automated. I mean, everyone wants an automated workflow where you can just hit a ‘Go’ button, and it can work from sample preparation to getting you an answer. Using just LC alone, it would be very difficult to do that. Adding SLIM technology to the workflow to make things more automatable is, in my view, going to drastically change analytical chemistry.

Then also, looking at the downstream part of analysis and the data that SLIM generates, it is much higher throughput; you see so many more things, and the quality and quantity of the data the instrument generates is unprecedented. But there is a bottleneck here in data processing. Data processing is already an issue today with LC-MS, but now you are increasing throughput so you have even more data. How do you alleviate that bottleneck? I think that automated data processing—using artificial intelligence and machine learning to pull insights out of these really complex datasets—is going to be huge.

With these algorithms, it is a case of “garbage in, garbage out.” But our data is very clean and very high-quality, and so I think we are well-positioned to be a leader here. With very high-quality data, the accuracy you can get out of these algorithms is going to be exciting. We are not quite there yet—you still have to build the instruments, generate the data, and train those algorithms properly to get these answers—but I do think our high-quality data will be extremely valuable to these future algorithm developers.