ALMOST a century ago, a chance discovery revolutionised medicine. Alexander Fleming left a petri dish of bacteria out while he went on a two-week holiday. On his return, he found that the dish had been contaminated by a fungus that produced an antibacterial substance. He named it penicillin, and it has since saved millions of lives.
Even in the modern world, drug discovery still essentially relies on chance. Pharmaceutical companies often screen thousands of compounds trying to find one with the desired effect.
The dream, though, is to make drug development a much faster and more rational process. Almost all the machinery of life is made of proteins, and pretty much every drug works by binding to proteins and changing what they do. In principle, it should be possible to use computers to work out the shape of proteins and then design drugs to bind to specific sites.
Thanks to the genome revolution, it is now easy to discover the DNA recipe for any protein. Figuring out their shapes, however, still requires expensive experiments that can take years. As a result, we know the recipes for 180 million proteins, but the shapes of only about 170,000.
That is about to change, thanks to artificial intelligence. DeepMind鈥檚 AlphaFold system solves protein shapes like a jigsaw puzzle. It works out the easy bits first, based on what it has learned about other proteins, and then gradually puts all the parts together over a matter of days. It is the first computer-based system to achieve results that match those of experimental methods 鈥� but much faster (see 鈥淒eepMind鈥檚 AI biologist can decipher secrets of the machinery of life鈥�).
The hope is that this kind of approach can be extended to predict how proteins interact with other molecules, to find
or design drugs that have specific effects. Accomplishing this would revolutionise medicine all over again.
Biologists will still have to carryout lab experiments and clinical trials. Biology is too messy and complicated for artificial intelligence to do all the work. But AlphaFold and its ilk should greatly accelerate the already astonishing pace of progress, and millions more lives could be saved.
