Deep learning software tools to the rescue of medical research3 min read . Updated: 10 Dec 2020, 10:13 PM IST
AlphaFold 2’s ability to predict how proteins fold could help us find cures for ‘uncurable’ illnesses
Move 37, five lines from the edge of the board, was a blinder. Startling and unexpected, it was proclaimed as a ‘clear mistake’ by watching experts. Lee Sedol, the opponent, flinched and was forced to think for 12 minutes before his rejoinder. An expert watching from the sidelines, Fan Hui saw through it, however: “It’s not a human move. I have never seen a human play this move. So beautiful. Beautiful. Beautiful. Beautiful." It was 2016, the game was Go, and the undefeated world champion Sedol was playing with a machine, AlphaGo, created by an artificial intelligence (AI) company called DeepMind. AlphaGo had defeated Lee in the first game, and this ‘inhuman’ move gave it a 2-0 lead. This was a Sputnik moment in AI, since Go, unlike chess, is an intricate, non-deterministic game. A game of instinct and feeling, it has 10 raised to the power 170 possible moves, more moves than “atoms in the known universe".
The force that AlphaGo unleashed is the power of deep learning, a subfield of machine learning built with algorithms inspired by neural networks of the human brain. DeepMind, the pioneer of this field, has used it to power face-recognition cameras and voice assistants, and defeat humans in various games. But its most astonishing application was unveiled last week: predicting how proteins fold.
“To understand life," according to The Economist, “you must understand proteins. These molecular chains, each assembled from a menu of 20 types of amino acids, do biology’s heavy lifting. In the guise of enzymes, they catalyse the chemistry that keeps bodies running. Actin and Myosin, the proteins of muscles, permit those bodies to move around. Keratin provides their skin and hair. Haemoglobin carries their oxygen. Insulin regulates their metabolism. And a protein called spike allows coronaviruses to invade human cells, thereby shutting down entire economies." Proteins are the origin of existence; the tail of a human sperm is composed of various proteins that work together to form a complex rotary engine that propels it forth to fertilize an egg and create life.
Proteins also do something else: they fold. The final intricate shape they take after folding determines their function. For instance, one of them can fold like “snakes in a can", which when embedded in a cell membrane, creates a tunnel that allows traffic in and out of cells. Other proteins form pockets called “active sites" that are shaped to bind—like a lock and key—a particular molecule, the spike on the coronavirus. By folding into myriad shapes, proteins that are made of the same stuff perform different roles. The Harvard Business Review draws an analogy: “All vehicles are made from steel, but a racecar’s sleek shape wins races, while a bus, dump truck, crane, or excavator are each shaped to perform their own unique tasks". If they fold wrongly, as they often do, they can cause horrific harm; the accumulation of misshapen proteins, for example, is said to cause Alzheimer’s, Parkinson’s, Huntington’s, and Lou Gehrig’s (ALS) disease. Wrong folds can cause cystic fibrosis and sickle cell anaemia, and other diseases that yet remain unknown.
Therefore, predicting how a protein will fold has been one of the big science challenges for decades. Scientists have used techniques like X-ray crystallography, but these are too slow. DeepMind’s learnings from AlphaGo, though, have been used to create AlphaFold 2, a deep learning program to predict how a protein will fold. Proteins are even more complex than Go; a protein could take any of as many as 10 raised to the power 300 different shapes. Go players do not play the game by knowing every step, but by taking short-cuts, often based on intuition. This is exactly how people played a game called FoldIt, a simulation of protein folding. DeepMind used this analogy to create AlphaFold. By feeding powerful computers earlier examples and patterns, developers taught them how to apply intuitive short cuts and rules-of-thumb. Much like Move 37, a computer can come up with solutions that stun human experts. In the latest test, a competition called CASP, AlphaFold 2 got a score of 92.4, way above anything ever before, placing a powerful tool in the hands of medical scientists. Knowing how proteins fold can help us address incurable-so-far diseases in an entirely new way. It could also prevent many illnesses.
As Andrei Lupas of Max Planck Institute exclaimed, “This will change medicine. It will change research. It will change bioengineering. It will change everything." One of the greatest ever mysteries in science has been unfolded.
Jaspreet Bindra is the author of ‘The Tech Whisperer’, and founder of Digital Matters