The Biosolutions Bulletin

Biology enters the age of design: Unlocking new realities and possibilities

• Novonesis • Season 11 • Episode 1

Share episode

Modern biology now relies on artificial intelligence, supercomputers, and advanced technologies to explore details once unreachable, uncover patterns invisible to the human eye, and design solutions that were once unimaginable. 

This audio article is part of The Biosolutions Bulletin. For the text version of this article and to get The Biosolutions Bulletin delivered monthly directly to your inbox, please go to: https://www.novonesis.com/en/biosolutions-bulletin

Biology, which is the study of all living things has evolved to become a transformative science creating biosolutions to tackle some of the world’s most pressing challenges. 

In this episode of the Biosolutions Bulletin by Novonesis we will focus on how the classic image of microscopes, petri dishes, and lab coats only tells part of the story. Modern biology now relies on artificial intelligence, supercomputers, and advanced technologies to explore details once unreachable, uncover patterns invisible to the human eye, and design solutions that were once unimaginable. 

When computers and AI first entered the biology lab 

How long do you think computers and AI have been part of biology? Ten years? Twenty? Thirty? Try seventy-five! 

It was 1946, and IBM had just unveiled the 602A, an electromechanical machine the size of a small desk, full of whirring gears and clattering switches. It was not a computer in the modern sense, but for its time, it was close to one. Scientists could feed it numbers, and in seconds it would spit out results that once took teams of researchers days to calculate by hand. 

By the early 1950s, machines like the IBM 602A had found their way into biology. In fields like applied genetics 1, researchers began leaning on these electromechanical workhorses to do statistical calculations. Suddenly, scientists were free to focus on thinking about what the numbers meant, rather than spending weeks just crunching them. 

And yet, these early machines were still just tools. They followed instructions blindly, without any “intelligence” of their own. 

That began to change in the 1960s. 

At Stanford University in the mid-1960s, a small team of chemists, computer scientists, and geneticists came together to tackle a problem that had puzzled researchers for decades: how to figure out the shapes of molecules. Molecules are the building blocks of chemistry and life itself, but at the time, determining their structure was an agonizingly slow process. Researchers relied on tedious manual calculations, sketches, and endless cross-checks, often spending weeks or months just to solve the shape of a single molecule. 

In 1965, the team built DENDRAL, one of the world’s first successful AI programs. DENDRAL was not “intelligent” in the way modern AI is, but it could mimic expert reasoning. Scientists taught it the dos and don’ts of chemistry, such as which bonds are stable, which are likely to break, and how molecules tend to behave. Using these rules, DENDRAL could test possibilities far faster than any human could. Where researchers once struggled for weeks, DENDRAL could narrow down the likely shapes in just a few hours. 

For chemists, it felt like a superpower. But it wasn’t just chemists who were intrigued. Biologists were watching closely, realizing that as their field moved deeper into the molecular world, they would soon face the same challenges. 

Still, DENDRAL and later systems like MYCIN and SUMEX-AIM, had their limits. These primitive AI tools were brilliant but strictly rule-bound. They could only do what experts explicitly taught them. As biology started generating higher volumes and complexity of information, these AI tools began to show their limitations, paving the way for a new generation of AI. 

Biology needed something different — systems that could learn from the data itself, spot patterns humans could not see, and make predictions we could not guess.  

Among all this complexity, one challenge stood out above the rest: understanding proteins. 

Solving the protein puzzle 

Inside every living cell, molecular biologists were uncovering an intricate world of tiny components: genes, DNA, enzymes, and metabolites. However, at the heart of almost everything they studied, they kept finding proteins. 

For most of us, the word protein brings to mind nutrition labels or fitness plans. It is just another food group for us, like carbohydrates or fats. But in biology, proteins are so much more than something we eat. 

Proteins are the molecular machines of life. They build our cells, repair damaged tissues, transport oxygen through our blood, send chemical signals between cells, and drive thousands of reactions that keep us alive. Every heartbeat, every blink, every immune response in your body — proteins make it happen. 

Ever wondered why we need to eat proteins in the first place? It is because proteins are built from smaller chemical units called amino acids — think of them as tiny LEGO blocks of life. There are 20 different amino acids, and by linking them together in different sequences, your body can build thousands of unique proteins, each with its own role to play. 

Understanding protein shapes became one of biology’s most urgent goals. But this turned out to be one of its hardest problems. For most of the 20th century, scientists relied on slow, painstaking techniques like X-ray crystallography, nuclear magnetic resonance (NMR), and mass spectrometry to determine these shapes.  

These methods are still used today because they produce incredibly precise data, but back then they were the only tools available, and solving the structure of a single protein could take months or even years. And yet, thousands of scientists around the world kept at it. Why? Because they understood something vital: knowing a protein’s shape unlocks many secrets to understanding life better and designing innovative solutions. 

To record all the discoveries with regards to the protein structure from across the world, in 1971, researchers launched the Protein Data Bank (PDB) — a global repository where scientists could deposit and access known protein structures. It started with just seven structures2. Over decades, thanks to improved techniques and thousands of dedicated researchers, the number grew steadily. By 2017, the PDB contained about 130,000 known protein structures3. 

It was an incredible achievement, but also a tiny fraction of what exists in nature. Scientists estimate there are hundreds of millions of proteins, and knowing their shapes is essential for solving some of biology’s biggest challenges. 

A breakthrough was desperately needed. 

When AI solved a 50-year-old protein challenge 

It was back in the 1970s that scientists discovered that proteins fold into intricate 3D structures inside our cells. 

Then, in 2018, something extraordinary happened. At Google DeepMind, researchers Demis Hassabis and John Jumper unveiled AlphaFold, an AI system unlike anything biology had ever seen. A problem that had taken scientists months or even years to solve for one protein, AlphaFold could now predict in just hours — and with astonishing accuracy. 

How? By learning the language of life. 

AlphaFold was trained on the vast datasets stored in the Protein Data Bank — decades of knowledge developed manually, hundreds of thousands of solved structures, and countless examples of how proteins fold inside a cell into 3D structures. But AlphaFold did not just memorize these shapes. It recognized patterns — the connection between a protein’s amino acid sequence and its structure, across the hundreds of thousands of proteins it memorized. It taught itself the hidden rules of folding, the subtle patterns and relationships that even expert biologists could not see. 

When AlphaFold was tested, the results stunned everyone. Give it an unknown protein sequence of amino acids, and it would predict its exact 3D shape — often with an accuracy matching gold-standard lab experiments that had taken years to complete. 

The impossible had now become a routine. 

By 2021, AlphaFold released predictions for over 200 million proteins — covering nearly every protein known to science. A challenge that had consumed half a century of human effort was transformed into something that could be solved in a single afternoon. 

The scientific world erupted. In 2024, Hassabis and Jumper were awarded the Nobel Prize in Chemistry 4 — a rare recognition not just of a biological discovery, but of artificial intelligence itself becoming an engine of science. 

And AlphaFold was just the beginning. Soon after, researchers at the University of Washington developed another powerful AI tool called RoseTTAFold, which can predict a protein structure in as little as 10 minutes on a gaming computer. RoseTTAFold goes a step further — it can also model complex biological assemblies 5 — essentially showing how multiple proteins fit, talk, and work together, like pieces of a molecular puzzle. 

With tools like RoseTTAFold, scientists can now simulate entire networks of protein interactions inside our cells, something that was impossible just a few years ago. Why does this matter? Because life itself runs on these networks. Proteins rarely work in isolation; they  rely on each other, bind together, and trigger chains of reactions that keep our bodies functioning.  

By understanding how proteins rely on each other inside our cells, scientists can now do things we once thought were impossible. They can create smarter medicines that target exactly the proteins that need support while leaving healthy ones untouched. They can predict how even a tiny mutation in a protein can have an impact across the network, sometimes causing diseases. And they can even design entirely new proteins with abilities nature never evolved, like enzymes that break down plastics. 

This marks a turning point in biology. Until now, much of science has been about observation — watching how life works and trying to understand it. But with these AI-driven tools, biology is stepping into an entirely new era: prediction. And prediction leads to design.  

Let's take a moment for some fun facts:

The genomics data explosion 

Since the first human genome draft was completed in 2001, genomics research has accelerated at an unprecedented pace. According to the NHGRI, 2 to 40 exabytes of genomic data are expected to be generated within the next decade. That is an almost unimaginable scale — one exabyte equals a 1 followed by 18 zeros. Managing and making sense of such massive datasets is far beyond human capacity. This is why AI and supercomputers are indispensable, transforming oceans of raw data into breakthrough insights that could power futuristic biosolutions.

Your genome is a biological hard drive

Sequencing just one human genome produces about 200 gigabytes of data — roughly the space needed for 200 copies of the movie Jaws! 

Genome sequencing biodiversity

The Earth BioGenome Project (EBP) aims to sequence the genomes of every known eukaryotic species on Earth — living things whose cells have a nucleus, such as plants, animals, fungi, and many microbes. That’s an estimated 1.8 million species! So far, scientists have successfully sequenced the genomes of 3,000 species from over 1,000 families. It’s a giant leap toward decoding life’s blueprint — but there’s still a long way to go!

Biological breakthroughs power biosolutions 

Understanding proteins and going a step further to predict their shapes is not just an academic breakthrough. It is the foundation of a new era where biology becomes a toolkit for solving some of humanity’s toughest challenges such as curing diseases, transforming agriculture, cleaning up the planet, and reshaping nutrition. 

For decades, biology was mostly about observation. But today, powered by AI, supercomputers, and revolutionary discoveries, we have crossed a threshold. 

We are no longer just watching life happen. By predicting protein shapes, modeling molecular interactions, and simulating entire biological systems, scientists are opening doors to solutions once thought impossible. 

And this is not a distant future. It has already begun. Here are three examples that show how this revolution is unfolding and what it could mean for our lives in the years ahead. 

1. Accelerating Drug Discovery 

By now, you know that proteins are at the heart of life. But here’s the other side of the story: proteins are also central to many diseases. 

Sometimes the problem comes from outside. A foreign protein enters our body through a pathogen, such as a virus or bacterium, and hijacks our systems. Other times, the problem comes from within. In genetic diseases like cystic fibrosis, Alzheimer’s, or certain rare metabolic disorders, our own proteins are missing, faulty, or misfolded, preventing them from doing their jobs. 

To treat these conditions effectively, scientists first need to identify the problematic protein, whether it is foreign or faulty. But identifying it is only the beginning. Designing a treatment requires knowing the exact 3D shape of the target protein and then creating a drug molecule that can bind to it perfectly, like a key fitting into a lock. Drug discovery has become faster, cheaper, and far more effective than ever before. 

We saw this transformation in action during the COVID-19 pandemic. Within days of the outbreak, scientists used advanced computing and AI to map the SARS-CoV-2 virus in extraordinary detail. Understanding the 3D structure of its spike protein was critical because it allowed researchers to design vaccines and treatments that could block the virus from entering our cells in the first place. It marked a milestone that would have been impossible without AI-driven biology. 

2. Personalized Medicine 

Right now, most medicines are designed with a one-size-fits-all approach. But in reality, your body is unique — and so are your proteins. 

AI is helping us change that. By combining insights from genomics (your DNA), proteomics (your proteins), and protein structure prediction, doctors are beginning to design treatments tailored to each individual. 

Take cancer therapy, for example. Tumors often grow because of faulty proteins that drive uncontrolled cell division. With AI, scientists can now identify the exact shape of those faulty proteins and match them with custom-designed drugs that target only the problem areas. AI tools are even being used to create “designer proteins” built specifically to fight cancer in certain patients, paving the way for truly personalized therapies6. 

This shift marks a turning point in medicine. We are moving away from mass-produced treatments and towards personalized solutions that are faster, more effective, and safer. 

3. Cleaning Up Our Planet 

Proteins are not just the building blocks of life. They can also become tools to heal the environment. Scientists have discovered enzymes that can break down plastics, but in their natural form, these enzymes work too slowly to tackle global pollution. 

In a groundbreaking research project published in Nature in 2020, scientists developed an enzyme that could break down plastic bottles made of Polyethylene terephthalate (PET) into reusable building blocks of plastic, known as monomers, within just 10 hours, an efficiency unmatched by natural enzymes7. This achievement points toward a future where plastics might be rapidly and more sustainably recycled. And the potential doesn’t stop there. Similar AI-driven methods are being explored to develop proteins to solve other environmental challenges, like capturing carbon dioxide from the air and cleaning up oil spills. 

Examples of some more exciting biosolutions on the horizon, powered by the fusion of biology, technology, supercomputers, and AI, include: 

Personalized Probiotics: Imagine gut-friendly bacteria designed specifically for you. These AI-tailored probiotics could rebalance your microbiome, improve digestion, strengthen immunity, and even influence mental well-being. 

Climate-Resilient Crops: Using AI-driven insights into proteins, scientists are engineering crops that can thrive in extreme heat, drought, and flooding. These super-resilient plants could help secure food supplies in a rapidly changing climate. 

Precision Gene Therapies: AI-powered tools are being developed to repair faulty genes safely and accurately, without altering the rest of your DNA. These therapies could unlock cures for diseases once thought untreatable. 

Biology is no longer just a science of observation. It is becoming a science of design. With the help of AI and supercomputers, we are starting to predict, model, and engineer life at levels once thought impossible. We are standing at the edge of a new frontier — one where biology, powered by intelligence both natural and artificial, could transform not just our understanding of life, but the very way we live it.  

Thank you for listening. This audio article is part of the Biosolutions bulletin by Novonesis. For the text version of this article and to receive the monthly Biosolutions bulletin directly in your inbox, go to: https://www.novonesis.com/biosolutions-bulletin.