Karl Krull, a PhD candidate working in Jeroen Krijgsveld‘s group at the DKFZ German Cancer Research Center in Heidelberg, is building innovative workflows that push single-cell proteomics towards dramatically higher throughput. By adding computational analysis to his repertoire, while focusing increasingly on the biological applications of his work, Karl is learning to uncover hidden patterns in his data that could answer fundamental biological questions.

The first fascination

Karl’s fascination with proteins began early in his career, with an internship at a pharmaceutical company, analysing intact antibodies using mass spectrometry. “This was my first contact with these fascinating machines,” Karl explains.

What began as technical curiosity evolved into something much deeper. The excitement of working with “solutions of hundreds of thousands of uncharacterised peptides” and piecing them together to identify distinct proteins captured his imagination. This passion led him from Munich to Heidelberg, where he found himself at the cutting edge of an emerging field: single-cell proteomics.

When Karl began his PhD, the ability to analyse the protein content of individual cells was just emerging as a possibility. “It was really just when I started that some labs were showing that this was actually feasible,” he says.

From methods to meaning

When Karl first entered the world of proteomics, his research began with a focus on methodology – developing and optimising workflows, increasing sensitivity, and preventing sample loss. “In the beginning, these technical aspects were really at the core of my work,” he reflects. “But over time, you’d begin to see all the possible applications on the horizon.” While Karl was immersed in the granular world of “tweaks to the workflows,” his focus broadened to include the biological applications as the technology matured. The transformation came through exposure. “The longer I am in the proteomics field, my perspective is changing,” Karl notes. “You’re getting in contact with new people, you’re networking, you’re seeing the perspective of other people.”

Now, Karl designs experiments with specific purposes in mind. The technical optimisation is no longer an end in itself, but a carefully calibrated approach to answering meaningful biological questions that can potentially impact clinical understanding.

This shift from methods to meaning has opened new avenues of discovery. Among the most significant findings is that cell populations are not homogeneous. “Populations have this heterogeneity,” Karl explains. “There are certain subgroups within populations that are primed for going into a particular developmental trajectory, while others prefer a different one.”

These discoveries about cellular heterogeneity have profound implications for understanding how tissues develop and how diseases progress. They also highlight the importance of studying cells individually, rather than in bulk.

The data challenge

As Karl’s research evolved, so did his relationship with data. Coming from a laboratory background, he initially had little interest in bioinformatics. “I remember talking to a bioinformatician who told me he lost interest in working in the wet lab and wanted to become a data scientist,” Karl recalls. “I never would have imagined undergoing a similar shift.”

But six months into his PhD, his perspective changed. “I realised, well, actually, this is quite interesting. Why not do more of this?” He took the plunge into programming, learning “the hard way” – something Karl would tell his younger self to do earlier.

This pivot to data science transformed not just his skill set but his entire perspective on the field. “I could never work only in the dry lab,” he says. “But producing your own data and working with it really is fantastic.”

The integration of laboratory expertise with computational analysis has enabled Karl to extract deeper insights from his experiments, uncovering patterns that might otherwise remain hidden in the data: “So many things you see in this data, you can’t even think of when you work with the samples in the lab. You might have an idea of what you’d like to show, but only when you work with the data do you realise all the gold that lies in it.”

When it comes to the writing process, however, Karl cautions not to get too caught up in the detail. “I think it’s important to take a step back. We want to squeeze out every little detail from our data. While this is important, when we write the story, it’s better to get a bird’s eye view and think about what serves the story versus what’s just additional analysis.”

Racing to catch up

Despite the remarkable progress in single-cell proteomics, Karl acknowledges that the field still lags behind other single-cell technologies, particularly transcriptomics (the study of RNA).

“Even if we use our fastest methods, we are still not analysing more than a couple hundred samples per week, which other single-cell technologies do easily within a day,” he explains. This throughput limitation presents a significant challenge when competing with more established techniques.

The technical hurdles are substantial. Current chromatography systems and mass spectrometry instruments, while increasingly sensitive, still require precious minutes to analyse each sample. Karl’s current research focuses on overcoming these bottlenecks to dramatically increase throughput.

His lab has embraced cutting-edge equipment to address these challenges, including the timsTOF and Astral mass spectrometer optimised with Aurora Series columns. “I think one of the strongest points of these columns, apart from the improved resolution, is the convenience,” Karl notes. “You have the emitter directly connected to the column, with no loose connection in between.”

The future: AI and multi-omics

Looking ahead, Karl sees two major frontiers for proteomics research. The first is the integration of multiple ‘omics’ approaches to study different aspects of the same cell. “This is truly remarkable, to think of possibilities that we can take two layers of the cell, and we can finally study the relations between these two layers.”

For example, combining proteomics with transcriptomics allows researchers to understand how many RNA transcripts give rise to how many copies of proteins. Similarly, Karl names the combination of metabolomics and proteomics as a potential avenue to learn about the activity profile of proteins.

The second frontier, which Karl predicts will also serve multi-omics analyses, involves artificial intelligence. Karl believes AI will transform how researchers analyse the complex networks of proteins within cells. “We know proteins go up, others go down, so we know a certain correlation between them,” he says. “We can study how groups of proteins are correlated within the same proteome.”

This would generate patterns ideal for AI analysis. “These correlations might be super interesting to be used with AI to make predictions about how these networks work,” Karl explains. “What’s the outcome if I change a certain aspect of this network and how this network rearranges to respond to the stimuli?”

AI approaches may prove particularly valuable for identifying biomarkers in clinical samples. Rather than focusing on individual proteins, AI can identify patterns and relations among dozens of proteins that, together, characterise a disease state—something that “is hard for conventional statistical methods to identify and hard for humans to grasp.” Karl’s group is using AI models on proteomics samples in their upcoming research – though we’ll have to wait for that research to be published before we know the exciting details.

The excitement of discovery

For Karl, perhaps the most memorable moment in his research career came when analysing single-cell data for the first time. “It was the moment when I analysed single-cell data, and you see these chromatograms, you see the first peak, and you think, ‘yes! We’re finally seeing something!’”

This sense of excitement continues to drive his work. As single-cell proteomics techniques mature, Karl and his colleagues are now poised to apply them to fundamental biological questions, collaborating with biologists and clinicians to explore the proteome’s role in development, disease, and cellular function, among the many other applications possible with this technology.

Would you like to be featured in an upcoming Community Newsletter? Email us at [email protected] and tell us a bit about yourself and your research.