Three researchers received the 2024 Nobel Prize in Chemistry for solving one of biology's toughest puzzles: figuring out the shape of proteins based on their chemical makeup.

John Jumper and Demis Hassabis of Google DeepMind developed an artificial intelligence system called AlphaFold that can predict protein shapes. David Baker of the University of Washington received the prize separately for related work on computational protein design. The Nobel Committee recognized their work as transforming structural biology — the study of how biological molecules are built — which for fifty years relied on expensive laboratory equipment and techniques to map these shapes.

The Problem

Proteins are chains of molecules called amino acids. The order of these amino acids tells you the ingredients in a protein, but not its actual three-dimensional shape — and that shape is everything. It determines what a protein does in your body.

For decades, scientists used sophisticated machines like X-ray crystallography and electron microscopes to figure out protein shapes, one protein at a time. It was slow and expensive. Researchers knew computers should theoretically be able to predict these shapes from the amino acid sequence alone, but the math was overwhelming. Even a modest protein has far too many possible shapes to check them all.

AlphaFold, trained on known protein structures, solved this. When tested against a standard science benchmark in 2020, it predicted shapes with the same accuracy as the old lab methods — but in seconds instead of months. That was unexpected. The team later released a public database predicting the shape of hundreds of millions of proteins, including virtually all proteins in the human body.

DeepMind then made it even easier: a simple website where researchers can upload a protein sequence and get a structure prediction back within minutes. No expensive computers needed.

What This Means

The Nobel Prize recognizes a machine learning system as a major scientific achievement — not just a helpful tool, but a genuine breakthrough in how we understand biology. That sends a message: artificial intelligence can do real science.

In practical terms, this speeds up drug discovery. When researchers want to design a medicine, they need to know how the drug molecule fits onto its target protein, like a key in a lock. AlphaFold lets them see that fit instantly rather than spending months or years in the lab.

There is another layer to this: once one hard problem is solved, the bottleneck moves elsewhere. Protein shape prediction is now the easy part. The harder questions are what proteins actually do inside cells, whether predictions work in real life, and how proteins bend and move over time. Those are the puzzles researchers are turning to now.