Science Chat with Dr. Daria Shlyueva
Dr. Daria Shlyueva is a Senior Scientist in Microphysiological Systems at Novo Nordisk, where she facilitates interactions between IT, data scientists, and biologists. She specializes in mapping and optimizing (meta)data capture and implementing sustainable process changes. Dr. Shlyueva received her PhD in Molecular Biology from the University of Vienna and her Master’s in Biomedicine from the Karolinska Institute. The views expressed here are Daria’s own and do not represent those of Daria’s employer.
Interview conducted by Siobhan Keegan, representing Watershed Bio.
Siobhan: What motivated you to pursue a career in science?
Daria: I knew that I always wanted to do something different, something other people didn’t understand very well, like genetics. For context, I got my high school education and Bachelor’s degree in Russia, where the system is quite different. I did a lot of math and physics coursework as a high schooler, but not as much biology. Eventually I decided to pursue biophysics for my Bachelor’s, which was similar enough to math and physics while adding in biology.
I never really considered industry or any other career path besides a PhD at first. With a PhD, it was very clear what the next steps would be: applications, interviews, and so on. It worked out in the end, and I don’t have any regrets about my decisions, even though I may not have thought through my choices very much at the time.
Siobhan: What advice would you give others looking to follow a similar path?
Daria: As a scientist, you really need to learn to tolerate uncertainty. Most of the time, things don't work as planned, and you still have to make decisions. In college lab classes, experiments are often well-defined and optimized with a lot of controls. It’s quite different from what happens in a real laboratory where you need to build experiments with controls yourself.
You also need to learn how to take negative feedback, as well as filter out the useful parts. As scientists, we are trained to criticize experiments, data, results… this is essential to science and how it keeps moving forward. But as a young scientist, it's very discouraging when someone criticizes the 10% of your experiments that actually worked out – you can’t even show 90% of your other attempts. It’s important to detach from your work and not identify with your results.
Siobhan: What else do you think is fundamental to teach future scientists?
Daria: First, we need to learn how to ask good questions. This is becoming increasingly important with the rise of AI. Right now, designing experiments has become sort of a data fishing expedition: let’s generate this dataset, and count on AI or ML to find something. But even if you have a properly trained AI, you still need to ask a specific, well-tailored question to get a useful response.
Second, we need to learn how to plan good experiments. Formulating a clear hypothesis or question and coming up with a set of controls and expectations to test whether the experiment worked is very difficult. In school, you’re usually given pre-designed, pre-planned experiments, which you’re just expected to execute and draw conclusions from. But in real life, you have to come up with your own question, and make sure you have proper positive or negative controls. Often, you want to test everything in one experiment, which is just not possible. There have to be compromises.
Siobhan: What would you consider a highlight of your research career?
Daria: One of my main motivations is to create work that is useful for other people. My PhD paper discussed how a specific hormone in fruit flies is able to activate different sets of genes in different cell types, which generated a lot of sequencing data. I had to upload this data into public repositories, which involved describing it in excruciating detail, or metadata. This was a lot of extra work, but I made sure to include enough information so that others wouldn't have to reanalyze it or figure out what was going on.
A few years later, I stumbled upon a paper that used this data to discover something completely new. To me, this is the epitome of good science. I was so proud that my data and metadata were good and clear enough for them to reuse, and ultimately allowed them to derive an entirely new discovery. Data from these repositories are frequently incomplete – you might get stuck preprocessing raw data, or you might not even be able to reproduce their results.
Siobhan: What was your experience like as a graduate student in Europe?
Daria: I would really recommend considering studying abroad in Europe. Your stipend is usually enough to live, and you have other benefits as well. In Vienna, I was able to have my own place relatively close to work, as well as health insurance and a good amount of vacation days. The public transportation was also great there, so I didn’t need a car.
If you’re worried about learning a new language, the working language at many institutions is often English. At the Vienna Biocenter, emails and other work-related communications were in English. Of course, when I went outside, people spoke German, but I was also offered free German classes, which allowed me to pick up enough to get by.
Siobhan: What do you work on now at Novo Nordisk?
Daria: I currently call myself a Metadata Engineer. I act as a sort of mediator or facilitator between biologists, software developers, IT professionals, and data scientists. When you’re performing an experiment in the lab, there is a lot of metadata that is not captured in the results. What exactly happened to the samples? What kits and instruments did you use? Who did the experiment, and when? This information is crucial for reproducibility and also for higher level analytics like machine learning. There are a lot of interesting features that can be found not in the data itself, but in the metadata.
Things don’t go perfectly in the lab – people are stressed, instruments don’t behave properly, and so on, and they just don’t have time to adequately capture all of the metadata. I’m working on developing a framework to integrate all the tools that we have and help scientists in the lab solve this problem. This requires not only a deep understanding of experimental processes, but also the ability to describe them to software developers and IT teams.
Siobhan: How did you find yourself in this niche role?
Daria: I think my experience in data science is foundational. If I had never written any code, opened a terminal, or connected to a server, it would be very difficult to understand all the steps necessary to do this work. My last role at Clade Therapeutics (recently acquired by Century Therapeutics), however, was really the key: I developed data models to show that it's possible to capture metadata in a very streamlined fashion with adequate templates and detail. It’s essential to build these steps into the experimental process rather than leaving people to write it down later.
Siobhan: What are the most challenging and exciting parts of your work?
Daria: I would say change in management can be quite difficult. My role doesn’t just involve process development – I also have to work with people to change their habits. When we introduce a new piece of software, how do we onboard people? How do we make sure that they're comfortable using it? If it turns out to be more tricky than we anticipated, should we choose another software?
On the flip side, I enjoy that my work requires a very diverse, unique skillset. No one specific person can teach me everything I need to know, and I like figuring it out myself.
Siobhan: Are there any books you would recommend to future scientists?
Daria: First, I really enjoyed Thinking Fast and Slow by economist Daniel Kahneman. He discusses our cognitive biases and how they influence our decisions, offering many examples. I think his main argument is that we need to activate what he calls our “slow thinking” system to make more logical decisions and avoid these biases.
Second, as a Master’s student I read Louis Pasteur, a book on the ethics of Pasteur’s research career by Patrice Debré. While Pasteur made significant contributions towards understanding how microorganisms cause diseases and how we can prevent this, he also made many questionable decisions leading to those breakthroughs. For example, he tested a rabies vaccine on a young boy without having enough animal data to support that it would cure him. He also faced a lot of criticism from animal activists for his work on animal research models. I think a lot of these problems still exist in science, and I enjoyed that the book isn’t black and white in its reasoning.