At ASMS 2015, GC-Orbitrap technology was unleashed onto an expectant analytical community. Here’s the back-story.
Historically, The Coon Research Group has been focused on protein analysis with mass spectrometry. More recently, we’ve been interested in small molecule work in the field of metabolomics. It’s pretty clear that quantifying small molecules can give a better correlation with biological phenotype than work further upstream. Moreover, until very recently, it was an area in serious need of new technology – and that’s where our interest in coupling gas chromatography with Orbitrap technology started. As a group, we’re very driven by new technology and its application to problems – especially when there’s such a fundamental gap. Sure, you can already detect these small molecules pretty effectively with mass spectrometry, but more often than not, you can’t understand their chemical formula. And it’s very hard to go from signals in a spectrum to biological function, if you don’t know what the molecule is… How can we identify these structures? Well, GC coupled with Orbitrap and its accurate mass capability seemed to be a great starting point to solve this problem.
Seize the gap
Clearly, there is a big difference between recognizing a gap and attempting to fill it. But fortuitously in the mid 2000s, we worked on a separate development project in collaboration with Thermo Fisher Scientific on electron transfer dissociation (ETD) for the Orbitrap, and we all recognized that it would be relatively straightforward to use that test system to try GC on an Orbitrap. The first ‘Frankenstein’s’ system certainly wasn’t practical, but it gave us data. In fact, it worked so well that another collaborative project was initiated to further investigate the potential. The short version of the story is that those initial efforts sparked Thermo Fisher Scientific’s development cycle (led on the R&D side by Brody Guckenberger and Scott Quarmby) for the commercial instrument that was released at ASMS 2015: the Q Exactive GCTM.
Of course, going from a proof-of- concept system to commercial instrument is in no way straightforward. And a big – often overlooked – part of the journey involves leveraging informatics. That’s where Nicholas (Nick) Kwiecien stepped up to the plate. We were generating a lot of data – and if you knew what you were analyzing, you could get the right answers. But how do you go backwards? Nick expressed interest in trying to figure it out and came up with some outstanding ideas on how to leverage accurate mass to get back to structure.
For the past 50 years or so, people have been using GC-MS systems equipped with unit resolution mass analyzers – and that means there are a lot great resources out there in terms of mass spectra repositories. The big question became: how can we leverage those resources? The answer led us to an innovative algorithm call high-resolution filtering (HRF), which is incorporated into the data processing software for the new instrument. HRF is uniquely enabled by the mass accuracy provided by Orbitrap technology and allows us to search existing reference databases with our acquired spectra in the same way as people have been doing for many years. But because we have such precise accurate mass, we can annotate every single peak in a spectrum using a simple combinatorial process. We take combinations of atoms from putatively identified molecules and map those forward to peaks. The approach was extremely discriminatory against false positives, and should really increase the throughput of mapping unknowns back to structure.
Taking GC Orbitrap for a spin
We’ve taken on a large number of proteomics studies – thousands of different cell lines or hundreds of tissue samples – to try to understand how protein abundance varies from sample to another. Now, we can complement all of those experiments with deep and high- quality metabolome profiles generated by the Q Exactive GC.
Our first acquisition of a 1200 sample set showed that the correlation between the metabolome and proteome profiles is remarkably close. It turns out that it’s much easier and faster to collect metabolome profiles GC-Orbitrap technology than it is to do proteomics. Given very large sample sets, we envision that our group – and many others – are likely to perform broad metabolome work to discover the most meaningful population subsets ahead of further work in the proteomics space.
With high quality data for both the proteome and the metabolome, you can investigate a small molecule with raised abundance and match it to the upregulated enzyme responsible. Such studies really allow you to understand function across the whole pathway at multiple molecular planes – from small molecule to protein.
The folks at ASMS 2015 that we’ve spoken to seem very interested in acquiring the technology, you can almost hear them thinking how they can integrate GC- Orbitrap technology into their work. And certainly there have been lots of questions. Perhaps more interestingly, people who have not traditionally done metabolomic work (certainly, not in the way that we have done) appear to be seriously tempted by the possibilities. Indeed, there is a distinct air of surprise surrounding some of the corresponding proteome and metabolome results we’ve been able to show – especially at the scale we’ve worked on.
In our own lab there have been moments of surprise too. Frankly, we were quite shocked by how well the new instrument worked right out of the box. We’d been using the proof-of-concept system, which was not really capable of the sample throughput needed for our large-scale studies. So when we set up the new instrument and realized that the crew at Thermo Fisher Scientific had taken the GC-Orbitrap concept to a completely different level. The Q Exactive GC was a real surprise – in a very good way. Suddenly, we had the throughput to match the quality of the data.
People also seem really excited about the capability of the software tools mentioned earlier that are included with the instrument. I think our most fundamental contribution (besides providing a motivating force for instrument development) is offering the solution to deal with the data. I guess that sort of capability is on everyone’s wish list – but previously we didn’t have the right data to permit those kinds of algorithms. Now, we do.
Our group is very excited about the instrument’s ability to map unknowns. But there are a lot of areas where scientists want to look for compounds that they already know – in pesticides and sports doping, for example. If you know what you’re looking for, the system still offers many benefits. The accurate mass really boosts sensitivity, because you can pick out targets from chemical noise. It means you can achieve the level of sensitivity for target analysis that is approximately the same as the most sensitive GC instrument – the triple quad. But (and it’s a big but) you can cover all the ions in the spectrum. Where sensitivity coupled with full scan capability is highly sought after, GC-Orbitrap technology will be of great interest.
From an informatics point of view, the fact that the data is so remarkably reproducible is also a pretty big deal. For our largest scale project to date are, we had to cope with data files that were collected 45 days apart – but the runs looked the same. Such reproducibility really helps you gain access to meaningful results much faster – and it also facilitates the writing of custom code to analyze your data.
At ASMS 2015, Orbitrap celebrated its 10th birthday. Where will GC-Orbitrap technology be at its own party in 2025? Well, you can bet that the instrument will continue to improve over the next 10 years – that’s just the trajectory of Orbitrap technology. At the same time, we’re rapidly going to get a handle on unknown mapping and quantitation. Assigning identifications to unknowns is the current bottleneck in metabolomics (and a lot of other small molecule analyses) – and that’s simply got to change. Accurate mass will allow people to go beyond current spectral libraries – and who knows how far software will have come by then? In terms of scale, today we’re running 1000 samples and that’s considered impressive. In 10 years, people won’t be shocked by numbers 10 or 20 times bigger. And at that scale, you can almost force discovery.
As the technology rolls out, it’s very likely that it will be used in areas that we cannot even envisage right now. Even talking to people at ASMS this year, exciting new ideas are already pouring forth; it’s clear that once you introduce powerful new technology, the sky is the limit.
This article was featured in the Analytical Scientist Publication by Thermo Fisher Scientific.