AlphaFold and the 2024 Chemistry Nobel

The protein folding problem

Proteins do most of the work inside living things. Each one is a chain of amino acids that folds into a precise three-dimensional shape, and that shape decides what the protein does. For decades scientists could read a protein's sequence far more easily than they could see its shape, which required slow and costly lab methods such as X-ray crystallography.

Predicting the fold from the sequence alone was one of biology's great open problems. Since 1994 a contest called CASP has tested prediction methods against structures that have been solved in the lab but not yet published.

What AlphaFold did

DeepMind, the London AI lab led by Demis Hassabis, entered CASP in 2018 with the first AlphaFold and placed first. In 2020 a redesigned system, AlphaFold 2, led by John Jumper, produced predictions that were often close to experimental accuracy. Organizers described it as a solution to a long-standing problem for many single proteins.

AlphaFold 2 is a deep learning system trained on known structures from the Protein Data Bank. It compares related sequences from many species to spot pairs of amino acids that tend to change together, a clue that they sit near each other, and it uses attention-based networks to refine a full 3D model.

A Nobel for computing and chemistry

In October 2024 the Royal Swedish Academy of Sciences awarded the Nobel Prize in Chemistry in two halves. One half went to David Baker of the University of Washington for computational protein design, building entirely new proteins. The other half went jointly to Hassabis and Jumper for protein structure prediction.

It was a striking year for AI at the Nobels. A day earlier, the physics prize honored John Hopfield and Geoffrey Hinton for foundational work on neural networks.

Why it matters to everyone

DeepMind and the European Bioinformatics Institute released a free database of AlphaFold predictions, which grew to cover over 200 million proteins, nearly every one known to science. Researchers use it to study diseases, design enzymes and plan experiments faster. A later version, AlphaFold 3, announced in 2024, extends predictions to how proteins interact with DNA, RNA and small molecules. Predictions still need checking in the lab, but they have changed where biologists start.

← ChatGPT's Launch and Record GrowthNeural Networks Explained →