鈥淚t has not escaped our notice鈥︹�� With those famous words published in 1953, James Watson and Francis Crick described the fundamental genetic significance of the double-helix structure of DNA, based on work by Rosalind Franklin. It was a pivotal moment in biology, allowing us to understand for the first time how living organisms store the recipes for making proteins 鈥� the molecular machines that do most of the hard work in our bodies 鈥� and pass them down the generations.
Another major step forward came in 2001, with the draft sequence of almost the entire human genome. That revealed the big picture, that we have around 20,000 recipes for proteins, but not the detail of what they do. To fully understand proteins requires knowing their three-dimensional structure. These can be obtained experimentally, yet that is often a very slow process. Alternatively, computers can be used to try to predict structures from the gene sequence alone, but that used to be very difficult.
Advertisement
Used to be. In November 2020, Alphabet-owned DeepMind revealed that AlphaFold, its AI system for predicting protein structure, could accurately do this for pretty much any protein from its gene sequence. Now, it has released predicted structures of nearly all the more than 200 million known proteins.
This is another huge step forward in biology. Protein structures are never going to get as much attention as, say, dramatic images from space, but knowing them can have a much greater impact on our lives, transforming medicine and perhaps also food, farming and synthetic biology.
While the Human Genome Project rapidly transformed research, it was a decade or so before it delivered practical benefits like better treatments. We can expect something similar with AlphaFold.
Much work remains to understand how protein structures relate to their functions and how proteins interact with each other and with other molecules. The hope is that this will now accelerate, hastening our understanding of how life works at the molecular level, of the universe within.
