When the idea of algorithms first came about in the 9th century, we had no idea of the many real-life problems that they would help to overcome. Some scientists believed that algorithms would remain purely theoretical, while others saw their potential applications in computation and engineering. However, none could have ever predicted the power of these algorithms to overcome our own blocks of understanding.
Let’s take the protein-folding problem, for instance. Back in 1972, Christian B. Anfinsen was awarded the Nobel Prize in Chemistry for decoding a fundamental principle that guides how proteins work. He figured out that, if you take a particular protein and denature it so that it loses its shape, once the process is reversed, the protein restores itself to the same exact configuration it had before. From this observation, Anfinsen deduced that all of the information needed to create a protein was contained within its peptide chain.
But how exactly does a protein know how to return to its original shape? This is an astounding mystery. Out of all possible versions or folds that a protein could manifest, it always returns to its most stable, or low-energy, version. This puzzle is what became known as the “protein folding problem,” a challenge that has been haunting scientists for over 50 years—but not anymore.
As of July 2022, AlphaFold, an advanced artificial intelligence algorithm created by Google’s DeepMind lab was able to predict the protein shape of over 200 million proteins, almost every single protein that exists in living things! This is a huge milestone in the field of molecular biology, and it’s changing how thought-leaders in the field are approaching the development of medicines.
The perfect example is David Baker, one of the leading researchers in protein design and a recipient of the 2021 Breakthrough Prize in Life Sciences, who, with the aid of AlphaFold’s profound findings, is engineering brand-new proteins! This kind of stuff is straight out of science fiction, yet it’s proving to be exactly what the field needed to outperform our current approach to medicine, and even tackle plastic pollution. Who knew that the use of artificial intelligence could cast such a wide net!
The most important thing to take away from this is that AI has already proven itself to be extremely helpful in performing computations at levels that far surpass our own, even though we have an exponentially larger capacity for computation (based on our neural networks). The challenge will be to ensure that it’s being applied ethically to solve our world’s problems.
And most notably, the data gathered by this algorithm, named AlphaFold (now in its 3rd incarnation), is freely available to the public!
As someone who is fascinated with how the world functions, this knowledge energizes me to continue exploring the intersection of AI and biology.