Metabolomics can now generate richer datasets than many labs can interpret. According to Dr. Matthew Lewis, Vice President of Metabolomics & Lipidomics at Bruker Daltonics, the field’s next step does not start with more data. Instead, it starts with stronger control over the data already being generated.
Asked what researchers should prioritize first, Lewis points to quality control. “To show that the approach is reproducible, that we can achieve the same results time and time again at different labs, requires a focus on quality. Quality is always first.”
That answer reflects a broader challenge in metabolomics. Discovery workflows often relax the strict criteria associated with targeted pharmacology or regulated assays. As more labs build their own methods, variation can enter through sample handling, preparation, instrumentation, data acquisition, and interpretation.
“In discovery mode, we relax all those criteria, and people start to do things their own way,” Lewis notes. “You start to get real heterogeneity in the approach and in the results.”
Working in analytical science?
Register for a FREE Separation Science account to subscribe to the Separation Science Newsletter.
Community efforts have helped bring reproducibility to the center of the discussion. Lewis points to initiatives such as the mQACC Consortium as evidence that quality now resonates across the field.
Building Confidence Before Analysis
Quality control starts long before data interpretation. Sample collection, storage, and preparation can all shape results before a vial reaches the instrument. Some of those steps fall outside an instrument manufacturer’s control. Bruker’s focus, Lewis indicates, starts with the fitness of the analytical system.
That includes the Qsee™ Performance Test, a software- and consumables-based QC workflow that uses a defined mixture of eight synthetic polymer compounds. A lab can run the mixture using predefined methods and use software to compare the resulting data against expected performance.
“The software will automatically chew on that data, reference it against what we expect, and tell you whether your instrument is good to go or whether there is a problem you need to tend to,” Lewis notes.
The same data can feed longer-term performance monitoring through Bruker’s cloud-based TwinScape™ platform. Users can compare system performance over time, against previous runs, or against a selected gold-standard dataset.
“We can look in depth at the performance characteristics of a system today versus yesterday, last week, or even against a gold-standard dataset,” Lewis adds. “We can also monitor measurement precision at the batch level.”
For metabolomics, these checks support a larger goal: giving researchers confidence that differences in the data reflect biology or chemistry rather than instrument drift, batch effects, or inconsistent performance.
Using TIMS to Improve MS/MS Data Quality
Ion mobility often enters metabolomics conversations through high-impact examples: two isomers separated, two compounds distinguished, a difficult pair resolved. Lewis views that framing as too narrow.
Bruker’s trapped ion mobility spectrometry (TIMS) can separate isomers and isobars, but Lewis argues that its routine contribution matters more than isolated examples. “We used to think TIMS was best exemplified by finding two isomers of a drug and showing that it can separate them,” Lewis reflects. “In reality, what TIMS does is much more powerful than that.”
In metabolomics, TIMS works throughout an analysis to reduce interference and improve MS/MS data quality. When two precursors overlap, conventional MS/MS can produce chimeric spectra—mixed signals that weaken automated annotation and lower confidence.
TIMS adds another separation dimension before fragmentation, helping clean up spectra before annotation tools interpret the data. “TIMS helps separate isomers, isobars, and interferences all the time,” Lewis notes. “It enhances the sensitivity and cleanness of our MS/MS spectra.”
That “cleanness” is important in discovery metabolomics, where annotation confidence depends on how well spectra match databases and how many interfering signals complicate that match. Lewis asserts that TIMS can also measure collision cross section (CCS), giving researchers an orthogonal separation and providing a characteristic measurement others cannot provide. “It is much more fundamentally important than separating the odd isomeric pair.”
How Chromatography Shapes Metabolome Coverage
For Separation Science readers, Lewis’s sharpest point may concern chromatography. Metabolomics often uses the language of untargeted analysis, but he challenges that premise. “There is no such thing as untargeted,” Lewis argues. “The second you select the column chemistry, you are targeted, whether you know it or not.”
Column choice shapes metabolome coverage. Stationary phase, polarity, ionization mode, and chromatographic strategy all introduce selectivity and bias. That bias may help capture one class of metabolites while leaving others underrepresented. “The choice of stationary phase imparts bias to the coverage of the metabolome achieved,” Lewis remarks. “That is often suffered but not often appreciated.”
Comprehensive coverage may require multiple approaches: HILIC in positive and negative mode, ZIC-HILIC, reversed-phase methods, ion pairing, GC–MS, NMR, or other complementary strategies. Each method opens some chemical space and closes off another.
Lewis views reversed-phase chromatography as powerful but imperfect for biological metabolomics. Other methods can reach difficult molecules, but they can also add complexity, variability, or practical burden. This creates a second problem: transferability.
Why Method Variation Limits Data Transferability
Metabolomics labs often build methods around local preferences, instruments, columns, gradients, and sample types. That flexibility can help individual labs optimize performance, but it can also weaken collaboration.
Lewis uses a memorable comparison. “Methods in metabolomics are like toothbrushes,” he notes. “Everyone has their own, and no one wants to use someone else’s.” The result: fragmented datasets, weaker cooperative database building, and less effective cross-lab annotation support. If every group uses a slightly different reversed-phase method, the community struggles to build shared resources that transfer across sites.
For Lewis, TIMS and CCS can help address that limitation. Chromatography remains powerful, especially because it occurs before ionization and supports quantitative rigor. But ion mobility can make data less dependent on one exact LC method. TIMS can operate with LC, flow injection, MALDI, or other workflows. That flexibility gives researchers another way to compare data across methods.
“Even if we use different chromatographic methods, we can match on mass, MS/MS, and now CCS values as well,” Lewis notes. “Chromatography plays less of a critical role in a relative sense, and data becomes more transferable.” That does not remove the need for careful chromatography. Instead, it adds another layer of information that can help bridge differences between methods, labs, and platforms.
Supporting More Reproducible Metabolomics
Metabolomics does not lack ambition. The field continues to push for broader coverage, faster analysis, stronger annotation, and better biological insight. But Lewis’s priorities point back to fundamentals: quality, reproducibility, and transferability. Better separation, larger libraries, and annotation workflows are all still important. Yet without standardized QC and methods that support cross-lab confidence, a dataset rich in information about the metabolome will have limited interoperability.
For separation scientists, the message is direct. Every chromatographic choice shapes the metabolome a lab can see. Every local method adjustment can affect how well another lab can compare, reproduce, or reuse the data. TIMS and CCS offer one route toward more transferable metabolomics by adding a measurement dimension that can travel across chromatographic differences. QC tools offer another route by helping labs understand when their systems can produce dependable data.
Together, those priorities point to a practical future for metabolomics: not more data for its own sake, but data that labs can trust, compare, and interpret with greater confidence.
