Understanding the structure of proteins—often referred to as the executors of life—is, in many ways, like solving the puzzle of life itself.

For decades, this challenge has been the subject of countless experiments and studies conducted by scientists around the world. Today, however, advances in artificial intelligence have brought the field to a new turning point. AI systems that once learned human languages are now deciphering the language of life encoded in proteins, fundamentally transforming the landscape of life science research.

At the center of this transformation is Professor Milot Mirdita. Recognized as one of the world's most influential researchers and named a 2025 Highly Cited Researcher (HCR), he has contributed to the development of ColabFold, a widely used protein structure prediction platform, while continuing to explore the fundamental mechanisms of life. Having joined Sungkyunkwan University School of Medicine this March, Professor Mirdita is now embarking on a new chapter in both research and education.

Let us hear his story.

| In addition to bioinformatics, you have also conducted research in machine learning and computer science. What motivated you to pursue bioinformatics as a research field?

My initial interest was actually much closer to programming, and for a long time biology was almost an excuse for me to do programming.

Of course, biology itself has now become the more fascinating part. There was no single moment that suddenly sparked my interest in biology. Rather, as I saw people actually using the software I developed, I gradually became genuinely interested in biology as a field.

In the beginning, what excited me most was simply the act of building something. Now, creating something with a purpose has become much more important to me. My ultimate goal is to develop tools that can make a positive impact on the world.

| You have conducted research in both Germany and Korea. Are there any similarities or differences between the two research environments?

Germany has a long-standing research tradition. Research environments at institutions such as the Max Planck Institutes and many of the country's historic universities are truly exceptional.

One of the greatest advantages I experienced at Max Planck was its international atmosphere. I had the opportunity to work with students and faculty members from many different cultural backgrounds.

That said, I am also very fond of Korea's research environment. It is highly dynamic and moving rapidly in a positive direction. Korea also invests heavily in research and development. To my knowledge, for quite some time Korea ranked among the world's leading countries in per-capita R&D investment, alongside Israel. For these reasons, I believe Korea is an excellent place to conduct research.

The students are also exceptionally talented. When I conducted research at Seoul National University, I worked in Professor Martin Steinegger's laboratory, which was filled with outstanding students. I hope to build a laboratory of a similar caliber myself.

| What led you to continue your research in Korea, and specifically at Sungkyunkwan University?

As I mentioned earlier, I found the unique strengths of Korea's research environment highly attractive. There are many excellent students, and the research funding system is quite strong.

Compared to systems in other countries that tend to be more bureaucratic, Korea's system is considerably more efficient and generally more accessible. Around that time, Professor Kyeong Kyu Kim, whose laboratory is next to mine, reached out to me directly, which ultimately led me to continue my research at Sungkyunkwan University.

The campus is beautiful, and I felt that it would provide an excellent environment for pursuing my research.

▲ Professor Mirdita during his time in the Steinegger Lab

| Bioinformatics may be unfamiliar to those outside the field. Could you briefly explain your area of research?

At its core, my research is about understanding proteins. I study not only proteins themselves, but also their functions and how we can manipulate them to improve human health.

Protein research is closely connected to artificial intelligence because algorithms allow us to analyze and compare proteins, enabling us to look deep into evolutionary history. In a way, it is like a time machine.

All living organisms are connected, and proteins provide one of the most direct ways to observe those connections. Even traces of common ancestors that existed billions of years ago can still be found in protein structures and sequences. While physical traits and genomes change relatively quickly, proteins often preserve much older evolutionary signals.

That is one of the reasons why I find bioinformatics such a fascinating field.

| You were recently recognized as a 2025 Highly Cited Researcher (HCR)*. Why do you think your research has received such widespread attention?

I believe one of the main reasons our research and software have been so well received is that we put a great deal of effort into making them easy to use.

Many research software projects are developed by a single student who deploys them on a server. Once that student graduates, the software is often left unattended. It is not uncommon for another student to spend weeks trying to get it running again, only to eventually give up.

Professor Martin Steinegger and I took a different approach. We invested significant effort into maintaining our software over the long term and consistently responding to users' questions. We wanted to make our research as accessible and usable as possible.

I believe that this long-term commitment is what ultimately enabled our work to have such a broad impact.

▲ 2025 Highly Cited Researcher (HCR)

Another project that has received considerable attention is ColabFold*. In many ways, ColabFold was made possible because the right people came together at exactly the right time.

We had already built much of the necessary infrastructure for other projects, although we did not anticipate that AlphaFold* would generate such a significant impact. Around that time, we learned that a researcher was experimenting with AlphaFold within the Google Colab environment. We realized that by combining that work with technologies our team had already developed, we could create a much more useful tool.

We worked intensively for several days and released the first version. In fact, we were able to launch ColabFold before DeepMind officially released code intended for general users.

▲ ColabFold Logo

ColabFold quickly gained widespread adoption and became a major success.

Today, it remains one of the easiest ways to predict protein structures. Users simply enter a protein sequence, click a button, and receive structure predictions within minutes.

The platform currently handles approximately 60,000 to 80,000 requests every day, making it one of the largest services of its kind, even when compared to other major servers.

* HCR (Highly Cited Researcher): An annual recognition by Clarivate that honors researchers whose publications rank among the most highly cited in their respective fields worldwide.

* ColabFold: A platform that enables users to run AlphaFold, an AI-based protein structure prediction model, quickly and easily through Google Colab, Google's cloud-based coding environment.

* AlphaFold: An AI-driven protein structure prediction model developed by DeepMind that predicts three-dimensional protein structures with high accuracy.

| Tools such as ColabFold and Foldseek that you helped develop have largely been released as open-source software. Was there a particular reason you chose this approach?

I strongly believe in open source.

Most of my research is funded through public resources, and I believe that the results of that research should be returned to the public as well.

Open-source tools are also highly accessible. I never wanted to restrict how people could use my software. When the code is publicly available, others can verify its validity, build upon it, and create entirely new things without needing my permission.

I think it is important not to control the direction of research myself, but rather to provide a foundation that allows others to create what they want. Ultimately, I want my software to be used as widely as possible.

| What do you hope to achieve through the development of these software platforms and research tools?

I hope people use these tools to conduct meaningful research.

Whether it is developing new therapies, advancing our understanding of life, discovering enzymes with biotechnological value, or conducting research that helps others, I hope these tools can contribute to a wide range of scientific endeavors.

At the same time, I would like to make researchers' lives a little easier. I want them to spend less time struggling with software installation issues and more time focusing on the research itself.

| Bioinformatics is often considered one of the fields most influenced by artificial intelligence. How do you think AI tools will affect protein structure prediction and the broader scientific landscape in the future?

I think this is a question for which no one can yet provide a definitive answer. It is something I discuss frequently at conferences and with fellow researchers.

Structural bioinformatics was not always considered a particularly popular field. Just a few years ago, I was even wondering whether I should leave protein structure research and move into genomics instead.

However, the emergence of protein language models and AlphaFold completely changed the situation. Protein structure prediction itself has now become a hypothesis that researchers can use to design experiments much more efficiently.

Many things that were technically possible before are now possible on a much larger and more ambitious scale.

Large language models (LLMs) present a more complex situation. They are incredibly powerful tools for writing code, but they still require substantial human oversight and verification.

More importantly, there is the question of how future researchers will develop expertise. In the past, expertise was built through reading papers, understanding code, and working through projects independently. AI removes part of that learning process.

As a result, I am not yet certain how people will develop deep expertise in the future. Nevertheless, AI is unquestionably a powerful tool, and I want to teach students how to use it effectively.

| You joined Sungkyunkwan University as a professor in the School of Medicine this March. As an educator and researcher, what goals would you like to achieve here?

First and foremost, I would like to build a strong research laboratory.

My goal is to create an environment where students can conduct research freely and grow as researchers. At the same time, I hope to continue producing impactful research.

I would like the resources developed in my laboratory to be widely used and ultimately benefit society as a whole.

In the long term, I hope to deepen our understanding of proteins and uncover how they interact with humans and other living organisms, contributing to improvements in people's lives through scientific discovery.

| Could you tell us more about your laboratory and the direction you hope it will take?

My laboratory is still in its early stages. We currently have three students and are gradually beginning new projects together.

I am primarily known for developing tools and methodologies for biological data analysis, and I plan to continue pursuing similar research directions in my laboratory.

At the same time, being part of a medical school opens many new opportunities. Although we are only at the beginning, I would like to collaborate with hospitals and establish new research projects.

There is also an abundance of fascinating data available within Sungkyunkwan University, and such data can open entirely new possibilities.

Currently, we are working on projects related to antiviral drug development, as well as projects aimed at more closely integrating systems biology and structural biology.

Ultimately, my goal is to create software that is easy for researchers to use, genuinely useful, and capable of making a meaningful impact. Through these tools, I hope to accelerate progress in biology and biomedical science.

▶ Mirdita Lab Website

| Lastly, do you have a message for Sungkyunkwan University students?

With the rapid advancement of AI, I believe we are living in a time that is both intimidating and incredibly exciting.

In times like these, gaining experience is one of the most important things you can do, regardless of your field. Whether it is a hobby, research, or any other pursuit, the things you enjoy and excel at will ultimately help you become a better person and a better researcher.

I hope SKKU students continue to stay curious, pursue their interests, and grow not only as researchers but also as individuals. Ultimately, well-rounded people can succeed in any era. I believe that if you approach everything with curiosity and an active mindset, while developing a deep understanding of your chosen field, good results will follow.

Interview: Jung Suyeon