🎤 Presenter: Caitlin Cain¹, Meriem Gaida², Pierre-Hugues Stefanuto², Jean-François Focant², Robert Synovec¹, Susan Jackels³, Kristen Skogerboe³ (¹University of Washington, Seattle, USA, ²University of Liège, Liège, Belgium, ³Seattle University, Seattle, USA)
💾 <a href="https://gcms.labrulez.com/labrulez-bucket-strapi-h3hsga3/O_15_MDCW_Uncovering_the_Volatile_Profile_of_Potato_Taste_Defect_in_Roasted_Arabica_Coffee_using_GC_MS_GC_GC_MS_and_Chemometrics_17c9ba7e27/O-15-MDCW-Uncovering-the-Volatile-Profile-of-Potato-Taste-Defect-in-Roasted-Arabica-Coffee-using-GC-MS-GCGC-MS-and-Chemometrics.pdf">PDF presentation</a>
<blockquote>
💡 Book in your calendar: <a href="https://gcms.labrulez.com/events/518">16th Multidimensional Chromatography Workshop (MDCW) 3 - 5. February 2025</a>
</blockquote>
<h3>Abstract</h3>
As coffee consumption increases globally, demands for specialty and/or high-quality coffee beans are primarily driven by consumer preferences toward the taste and aroma of the final brew. However, the presence of chemical defects can negatively impact the sensory properties of coffee. Potato taste defect (PTD), unique to East African coffee, is characterized by a distinct musty, vegetable, raw potato aroma due to the presence of 2-isopropyl-3-methoxypyrazine (IPMP). This work develops a volatile fingerprint of PTD in roasted arabica coffee using headspace-solid-phase microextraction-comprehensive two-dimensional gas chromatography with time-of-flight mass spectrometry (HS-SPME-GC×GC-TOFMS) and chemometrics. 
Tile-based Fisher ratio (F-ratio) analysis discovered over 300 volatiles that discriminated clean (i.e., no off-odor) coffee samples from those identified as having a strong PTD odor (p-value less than 0.01). Analytes associated with unpleasant sensory descriptions were positively correlated with the odor of PTD, while analytes associated with desirable aromas were found to be negatively correlated with odor severity. Principal components analysis (PCA) highlighted that the coffee samples clustered based on the presence of PTD. Furthermore, use of the class-distinguishing analytes discovered by F-ratio analysis for partial least squares (PLS) regression produced accurate models of IPMP concentration with low prediction errors (~ 10 % error). Investigation of the PCA and PLS loadings revealed the volatile compounds which are most important to predicting PTD in roasted coffee. Ultimately, these results reveal the complex volatile signature of PTD in East African coffee beans, demonstrating that the defect has outcomes that extend well beyond the concentration of IPMP.

Sensory-classified roasted arabica coffee with GC×GC-TOFMS and chemometrics (Caitlin Cain, MDCW 2024)

Dr. Caitlin Cain discusses her groundbreaking research on potato taste defect (PTD) in coffee using comprehensive two-dimensional gas chromatography (<a target="_blank" rel="noopener noreferrer" href="https://gcms.labrulez.com/library?instrumentationItems=2">GC×GC</a>). Learn how this advanced technique revealed 359 class-distinguishing analytes related to PTD, providing unprecedented insights into this complex flavor issue.<h4>Key points covered:</h4><ul><li>Advantages of GC×GC over traditional one-dimensional GC for complex samples</li><li>Optimization of headspace solid-phase microextraction for coffee volatile analysis</li><li>Surprising findings from the comprehensive volatile profile of PTD-affected coffee</li><li>Implications for coffee farmers and the industry in combating PTD</li></ul>Discover how this research is advancing our understanding of coffee chemistry and potentially improving coffee quality worldwide.<h3>Video Transcription</h3><h5>Interviewer: Caitlyn, congratulations on being recognized as a Rising Star in Separation Science. How did it feel to receive the Outstanding Achievement in Gas Chromatography award?</h5>Caitlyn: Thank you! It was a huge honor. The award essentially recognized the full five years of my PhD in GC. I’ve since moved from GC into LC for my postdoc, so seeing my PhD work acknowledged like that was really special. I cared deeply about those projects, and it was gratifying that others saw their value.<h5>Interviewer: Had you come across Organomation or used nitrogen evaporators before this?</h5>Caitlyn: No—this was my first exposure to the company. In my GC sample prep we didn’t do much with evaporation-based extractions, so the award actually prompted me to learn more about that side of sample preparation.<h5>Interviewer: Big-picture question: it feels like many applications are moving from GC to LC. What’s your take?</h5>Caitlyn: I don’t see it as LC “replacing” GC. Both are excellent but target different application spaces. GC shines in areas like petrochemical/fuels and many food analyses; LC is great for metabolomics and even proteomics. I’m passionate about bringing GC and LC together—looking at volatile and nonvolatile sides of the same problem.<h5>Interviewer: What initially drew you to analytical chemistry and chromatography?</h5>Caitlyn: I’ve always loved chemistry—my mom is a high-school chemistry teacher. As an undergrad, I worked with Dr. Saron and Dr. Maryann Collinson on LC stationary-phase gradients and column modifications. They treated me like a grad student: my own project, my own milestones. Watching peaks emerge and seeing a modification succeed (or not) hooked me on chromatography. That led to my PhD and now my postdoc.<h5>Interviewer: Advice for early-career scientists who want to contribute meaningfully in analytical chemistry?</h5>Caitlyn: Zoom out and connect ideas. It’s fine to focus on an application or a single technique (GC or LC), but ask what concepts transfer across techniques. A lot of exciting science happens at the intersections. And embrace failure—I have plenty of bad ideas. Treat setbacks as learning opportunities on the twisty research path.<h5>Interviewer: What innovations might shape separation science over the next few years?</h5>Caitlyn: We’ll keep chasing higher resolution and peak capacity. On the instrument side: new column chemistries, moving to smaller-ID/capillary formats in LC, and on the GC side further chemistry refinements and multidimensional separations (a major focus in my PhD). On the computational side: non-targeted analysis, stronger math and statistics to extract meaningful analytes, and growing roles for machine learning and AI—areas I’m beginning to explore.<h5>Interviewer: Let’s talk about your 2024 publication on “potato taste defect” (PTD) in coffee. How did that project start?</h5>Caitlyn: Through collaboration. An ethical coffee roaster sourcing from African regions asked researchers at Seattle University (a primarily undergraduate institution) to investigate. They began by quantifying a known PTD biomarker—2-isopropyl-3-methoxypyrazine (IPMP)—and quickly saw coffee’s complexity: hundreds of peaks, with hints that more than IPMP might be changing. They then approached my PhD lab at the University of Washington because we specialize in comprehensive two-dimensional gas chromatography (GC×GC) and non-targeted discovery.<h5>Interviewer: For readers new to GC×GC—how does it improve on one-dimensional GC for complex matrices?</h5>Caitlyn: In 1D GC, separation is mainly by boiling point on a single column before detection (MS, FID, etc.). In GC×GC, effluent from the first column is periodically “sliced” by a modulator and re-injected into a second column with complementary selectivity. The continuous 1D→2D process boosts selectivity and greatly increases peak capacity—turning unresolved 1D “humps” into resolved features in 2D space.<h5>Interviewer: How did you optimize headspace SPME for coffee volatiles, and what were the hurdles?</h5>Caitlyn: Seattle University had an initial HS-SPME method tuned for IPMP, but it required 60 minutes at low temperature—not ideal when you want to run hundreds of samples by GC×GC. I balanced analyte signal against throughput by optimizing extraction temperature and time. By increasing temperature and halving extraction time, we maintained comparable analyte signal while improving overall throughput.<h5>Interviewer: Your study reported 359 class-distinguishing analytes related to PTD. What stood out most?</h5>Caitlyn: Three things:<ol><li>We confirmed IPMP changes between PTD and clean coffee—but found ~358 additional volatile features shifting as well.</li><li>Roughly 300 analytes showed higher signal in clean coffee and dropped in PTD samples. Many of these (e.g., pyrazines, furans) drive desirable aroma—cocoa, nutty, roasted notes. Their depletion in PTD samples erases pleasant flavors, which amplifies the earthy, moldy, “potato-like” perception.</li><li>Among analytes elevated in PTD samples, we saw more phenols and phenol derivatives formed during roasting (often linked to chlorogenic acid degradation). Other coffee studies suggest microbial activity can raise chlorogenic acids in green beans; after roasting that could explain higher phenolic products—potentially connecting microbial infestation with the PTD profile we observed.</li></ol><h5>Interviewer: Why does potato taste defect matter?</h5>Caitlyn: Nobody wants coffee that tastes like potatoes—and with PTD you also lose the good flavors. It’s a double hit: the off-note increases while positive aroma compounds decline.<h5>Interviewer: How can your findings help farmers and the coffee industry?</h5>Caitlyn: PTD is under-studied; causes and mechanisms aren’t well established. Our comprehensive volatile profile shows what’s happening in the beans, giving farmers and roasters a data-driven baseline to develop interventions. Ideally this helps direct resources toward mitigating the Antestia bug (linked to PTD in some regions) and toward biochemical studies to connect insect damage to volatile outcomes. Ultimately, the goal is practical solutions that protect quality and help African farmers command better prices.<h5>Interviewer: Anything we didn’t cover?</h5>Caitlyn: That’s the gist. GC×GC was a particularly powerful tool here. And thank you again—to you and to Organomation for sponsoring the award.<h4>Key takeaways</h4><ul><li>GC and LC are complementary, not competing; integrating both offers fuller chemical insight (volatiles + nonvolatiles).</li><li>GC×GC dramatically increases peak capacity and resolves complex matrices like coffee headspace.</li><li>Optimized HS-SPME (higher T, shorter time) maintained signal while improving throughput.</li><li>PTD impacts hundreds of volatiles: it removes desirable aromas (pyrazines, furans) and raises phenolic products—consistent with microbial influences and roasting chemistry.</li><li>Future drivers: new column chemistries, multidimensional separations, and ML/AI for non-targeted discovery.</li></ul><blockquote>This text has been automatically transcribed from a video presentation using AI technology. It may contain inaccuracies and is not guaranteed to be 100% correct.</blockquote><blockquote><h4>Concentrating on Chromatography Podcast</h4>Dive into the frontiers of chromatography, mass spectrometry, and sample preparation with host David Oliva. Each episode features candid conversations with leading researchers, industry innovators, and passionate scientists who are shaping the future of analytical chemistry. From decoding PFAS detection challenges to exploring the latest in AI-assisted liquid chromatography, this show uncovers practical workflows, sustainability breakthroughs, and the real-world impact of separation science. Whether you’re a chromatographer, lab professional, or researcher you'll discover inspiring content!You can find Concentrating on Chromatography Podcast in podcast apps:<ul><li><a target="_blank" rel="noopener noreferrer" href="https://open.spotify.com/show/1wZg67rCjoGju4yhqxI3tL">Spotify</a></li><li><a target="_blank" rel="noopener noreferrer" href="https://podcasts.apple.com/tz/podcast/concentrating-on-chromatography/id1788429185">Apple podcast</a></li><li><a target="_blank" rel="noopener noreferrer" href="https://podscan.fm/podcasts/concentrating-on-chromatography">podscan</a></li></ul>and on <a target="_blank" rel="noopener noreferrer" href="https://www.youtube.com/@ChromatographyTalk/videos">YouTube channel</a></blockquote>

Unraveling Coffee's Potato Taste Defect: Insights from 2D Gas Chromatography

🎤 Connor JohnsonConnor Johnson, a researcher from the University of Alberta, discusses his award-winning honours project analyzing the volatile organic compounds (VOCs) in two banana species using headspace gas chromatography-mass spectrometry (HS-GC-MS). &nbsp;He completed this specific project as an undergraduate at Thompson Rivers University (TRU).For over 60 years, commercial banana flavoring has remained unchanged—even though the fruit it's supposed to mimic changed in the 1950s. Connor's research reveals why fake banana tastes fake: the commercial banana extract contains only 3 compounds compared to 18+ in real bananas, missing critical compounds that create authentic banana flavor.<h4>This episode covers:</h4><ul><li>The history of banana flavoring and the myth of the Gros Michel banana</li><li>What Connor discovered when comparing Cavendish vs. Gros Michel bananas</li><li>The real compounds behind authentic banana flavor (hint: it's not just isoamyl acetate)</li><li>Why headspace GC is ideal for volatile organic compound analysis</li><li>Challenges with sample prep and instrument troubleshooting in research</li><li>How this research could revolutionize flavor chemistry in the food industry</li><li>The broader applications of comparing artificial flavorings to real fruits</li></ul>Connor won two national conference awards for this work and shares insights into the analytical challenges of flavor chemistry, including instrument downtime, sample matrix effects, and why creating authentic synthetic flavoring is harder than it seems.Perfect for chemistry students, flavor scientists, and anyone curious about why banana candy tastes nothing like real bananas.<h3>Video Transcription</h3><h4>Why banana flavor doesn’t taste like bananas</h4>Connor: We have to go back about a year and a half. I was working on an organic project that was unrelated, but I had a supervisor I really wanted to work with. I asked if she would supervise an honors project. She said yes and asked what I wanted to work on.I said it would be really cool to do something with food science. Then I asked: “Why does banana flavoring not taste like bananas?” That was the original big-picture question. She’s an organic chemistry professor—Jess Allingham from TRU—and she said, “Let’s do it.”My proposal even said I was going to try to make a banana flavoring, but first I had to figure out what was in bananas. So the first step was analysis, and that ended up becoming the whole project.We got to the point where we found the different VOCs we could detect in banana species and in banana flavoring, and we compared them. Technically, you could make a new banana flavoring—but it’s expensive to make many different esters or alcohols, and you’d also need food-safe production conditions, which we didn’t have. So that part became more of a long-term “someone else could do it later” idea.It became a passion project. People love the banana project. Everyone eats food, so we should know what’s in it.David Oliva: That reminds me of banana candy. I enjoy it, but it doesn’t really taste like an actual banana.Connor: Exactly. It resembles banana, but it doesn’t truly taste like it. Fruits do their job of “making flavor” really well. Banana candies with actual banana are hard to do—people have tried with varying success.<h4>Comparing banana species and “the banana flavoring myth”</h4>David Oliva: You analyzed two different banana species. What compounds did you identify, how did they differ, and what surprised you?Connor: The most important compound—off the top of my head—is isoamyl acetate (the classic “banana candy” compound). It was the largest-abundance compound, about 30% of the profile in both.The two bananas I analyzed were:<ul><li>Gros Michel (often discussed as the older “classic” banana)</li><li>Cavendish (the typical grocery-store banana)</li></ul>I also looked at a commercial banana flavoring (I won’t name the brand).There’s a common idea that banana flavoring is modeled after the old banana. It’s not. That’s a myth.A quick history note: In the 1930s, Gros Michel struggled due to Panama disease, and bananas became difficult to get. Banana flavoring became popular as a substitute—often relying heavily on isoamyl acetate because it tasted the most “banana-like.” People later started saying the flavoring was modeled after the old banana, but the overall takeaway is that the banana species are actually very similar in their flavor compounds.There were minor differences in isoamyl acetate:<ul><li>Gros Michel: about 36%</li><li>Cavendish: about 33%</li></ul>Across both bananas, the compounds were relatively the same. If there were differences, they were often “fingerprint-level” amounts—but even small differences can affect flavor. Also, sugar content and subtle compositional differences affect how you perceive the taste. Gros Michel is often described as a dessert banana, and it’s sweeter than Cavendish.David Oliva: So isoamyl acetate is the primary “banana flavor” compound, but other minor compounds also shape the authentic banana profile—along with sugar content?Connor: Yes. It’s a combination of the relative amounts of compounds and the sugar content influencing perception.<h4>Banana diversity and real-world sampling challenges</h4>David Oliva: You mentioned different banana varieties in other regions.Connor: There are thousands of banana species. In some regions you can find bananas with unusual colors and distinct flavor notes. We only “know” the variety that ships well—Cavendish.For the project, sourcing Gros Michel was expensive. We imported bananas via a distributor out of Miami. Shipping was about $100, and it ended up being around $20 per banana in Canadian dollars. They also spoil quickly. If you get a chance to try a Gros Michel banana, it might not change your life—but it might.<h4>Instrument downtime and what Connor would do differently</h4>David Oliva: It sounds like instrument downtime impacted your data collection. If you could run the project again, how would you do it differently?Connor: I wish I had known the issue earlier. The problem was the headspace injector needle (or something in that pathway). Some days I’d run Cavendish samples and get great peaks. The next day—same parameters—I’d get only baseline. That happened for months. People changed the column and other parts, but it ended up being the headspace needle.If I had known sooner, I would’ve had more time to run samples and optimize parameters. My parameters worked to detect the VOCs, but I would’ve liked better separation and shorter run times so I could run more samples.At TRU we had one GC-MS that also had to be used for classes and other work—having a dedicated instrument is a dream for many students. Luckily, it was fixed in the last few weeks before the report was due, so I got results and finalized the project.<h4>Headspace GC-MS workflow for banana VOC profiling</h4>David Oliva: Walk us through the headspace GC workflow, including sample prep steps, and why you chose headspace over direct injection or other methods.Connor: The logic for headspace was: I read that the VOC profile (what you smell) is closely related to the flavor profile. There wasn’t anything directly comparing the banana species I chose using headspace. Also, early on I was still thinking about potentially creating a new flavoring and wanted to focus on VOCs.Sample preparation (final workflow):<ul><li>Weigh out about 0.5 g banana into a headspace vial</li><li>Mash the banana</li><li>Add a small amount of water (about 1 mL, mainly for standardization)</li><li>Place the vial in a warm water bath at about 50 °C for 5 minutes to help volatilize lower-boiling compounds</li><li>Run the vial on the GC-MS using headspace injection</li></ul>I don’t remember every parameter now, but the headspace oven time was not very long because I pre-volatilized it. I didn’t want to “ripen” the banana during heating, because ripening can generate unwanted VOCs. The goal was to profile what a banana actually tastes like—not a highly ripened or processed banana.Compound identification and data handling:<ul><li>I used the NIST library/database to identify compounds (this was a qualitative study)</li><li>I was not quantifying concentrations; I used peak area percentage (relative composition) to compare samples</li></ul>David Oliva: You’re the second researcher in a row who’s talked about the value of NIST.Connor: It’s great. It saved a lot of headache. If I could’ve made it quantitative, I would have, but that led to additional issues.<h4>Extraction attempts and why headspace remained the core method</h4>David Oliva: Did you use extraction techniques like SPME, and how did that impact sensitivity and reproducibility?Connor: Before I even started analyzing by headspace, I tried to extract VOCs from the banana material. I tried:<ul><li>Steam distillation (but heating to ~100 °C effectively “ripens/cooks” the banana and alters the VOC profile)</li><li>A small-scale CO₂ extraction concept using dry ice and warming to drive CO₂ through the sample (great idea, but bananas are dense and high-water-content; it didn’t penetrate well)</li></ul>I tried many variations: changing vessel size, different sample prep, freeze drying, nitrogen freezing, drying and grinding—many methods. I couldn’t reliably extract VOCs that way. Maybe solid phase extraction or supercritical fluid extraction could work, but it wasn’t successful for me.I considered SPME, but stayed with headspace to see results. Looking back—with more GC knowledge now—I wish I had run a comparison between SPME headspace and static headspace to evaluate sensitivity differences and how much analyte might be missed.<h4>Matrix effects and detection challenges in fruit samples</h4>David Oliva: What detection challenges did you face with the banana matrix?Connor: Any fruit has matrix effects. Because I was working with solid fruit, it wasn’t ideal. I didn’t address matrix effects in the project, but I wish I had done something like standard addition to quantify how much signal is suppressed or altered.At the time, I was more focused on getting detection at all than on correcting matrix effects. It’s a shortcoming I’d like to revisit in the future—maybe later in my career.<h4>From Headspace GC to LC-MS instrument development</h4>David Oliva: You’re now doing LC-MS instrument development. What lessons carried over from your Headspace GC work?Connor: Being comfortable working with instruments was a big advantage. At a smaller university, I got more hands-on time than many undergrads. That comfort translated well into LC-MS work.I can’t share details of what I’m working on now, but I’m working on a new interface. The confidence from hands-on GC experience helped me feel more comfortable with software and instrumentation. A lot of people are hesitant to approach expensive instruments, but we still need people who can work with instruments—not only analyze data.<h4>Advice for students moving into separation science and MS</h4>David Oliva: For undergraduates starting out—especially those transitioning between techniques—what advice would you give?Connor: Have the initial conversation. Say: “I really like the work you’re doing. I want to learn more about LC-MS or GC-MS or ICP-MS. I want to work with you—can you help me?”That first email or meeting is the hardest step for many people. Showing genuine interest gets your foot in the door and can lead to longer conversations and opportunities.<h4>Future directions and applications in flavor science</h4>David Oliva: Anything else you’d like listeners to know?Connor: If anyone wants to work more on the project, feel free to reach out. This kind of work is applicable to many foods: compare grape flavoring vs. grapes, cherry flavoring vs. cherries, strawberry flavoring vs. strawberries.It could also be used for adulteration testing: checking whether a “natural flavor” is truly natural or essentially a small set of compounds. That was one takeaway—some flavorings claim additional natural flavors but may contain only a few compounds.There’s a lot of work left in the flavor industry. If you’re interested, dig into it. If you ever think “why does this taste weird?”—there’s probably a compound behind it worth studying.David Oliva: I was attracted to your work because it’s a question everyone has asked—bananas vs. banana flavoring. Thanks for coming on and sharing it.Connor: Thanks for the invite. Maybe in the future I’d be happy to come back and talk about more GC work when I have a new project—or when I’m a PI.David Oliva: Awesome. Thank you.<blockquote>This text has been automatically transcribed from a video presentation using AI technology. It may contain inaccuracies and is not guaranteed to be 100% correct.</blockquote><blockquote><h4>Concentrating on Chromatography Podcast</h4>Dive into the frontiers of chromatography, mass spectrometry, and sample preparation with host David Oliva. Each episode features candid conversations with leading researchers, industry innovators, and passionate scientists who are shaping the future of analytical chemistry. From decoding PFAS detection challenges to exploring the latest in AI-assisted liquid chromatography, this show uncovers practical workflows, sustainability breakthroughs, and the real-world impact of separation science. Whether you’re a chromatographer, lab professional, or researcher you'll discover inspiring content!You can find Concentrating on Chromatography Podcast in podcast apps:<ul><li><a target="_blank" rel="noopener noreferrer" href="https://open.spotify.com/show/1wZg67rCjoGju4yhqxI3tL">Spotify</a></li><li><a target="_blank" rel="noopener noreferrer" href="https://podcasts.apple.com/tz/podcast/concentrating-on-chromatography/id1788429185">Apple podcast</a></li><li><a target="_blank" rel="noopener noreferrer" href="https://podscan.fm/podcasts/concentrating-on-chromatography">podscan</a></li></ul>and on <a target="_blank" rel="noopener noreferrer" href="https://www.youtube.com/@ChromatographyTalk/videos">YouTube channel</a></blockquote>

Why Fake Banana Flavor Doesn't Taste Like Real Bananas

Article

Product

Events

Webinars

Application

Academy

Video

Poster

Scientists

Plastics

Environment

Food

Water

Health

Presentation

Scientific article

Interview

Toxicology

Popularization

Science and research

Software

Covid-19

Career

Laboratories

Various

Zaměření

Type of News

Focus

The Multidimensional Chromatography (MDC) Workshop draws experts in this exciting field to share and discuss their research.

The Multidimensional Chromatography Workshop

Organomation, founded in 1959, designs and manufactures high-quality nitrogen evaporators and extraction systems. Known for innovation and durability, their lab instruments are used globally for efficient sample preparation with strong customer support.

Why Fake Banana Flavor Doesn't Taste Like Real Bananas

Unraveling Coffee's Potato Taste Defect: Insights from 2D Gas Chromatography

Sensory-classified roasted arabica coffee with GC×GC-TOFMS and chemometrics (Caitlin Cain, MDCW 2024)

Upcoming events

Integrated Analytical Workflows for Better Performance

LECO Training Course Pegasus HRT+ / HRT+ 4D

Agilent Lab Innovation Day for Academia

Webinars

Advances in AFM for the Characterization of 2D Materials

Mastering Gas Sampling: Valve Fundamentals to ASTM Ready RGA & TOGA Analysis

Column Confidence: Making the Right Choice for Your GC Needs