How is artificial intelligence helping us learn about the body?
From the doctor
Biomedical science typically proceeds in small steps, with very occasional breakthroughs. Artificial intelligence has been changing this pace dramatically, most notably with recent (and astonishing) progress in helping us understand important body processes. Before explaining, I need to set the stage.
The human body is composed of cells — trillions of them — of many different types. For example, the cells in your eyes that see light are different from the cells in your stomach that make acid. The function and behavior of each cell is governed by tiny "machines" within it.
Those machines are made of molecules (including nucleic acids, ions and, above all, proteins), each with its own shape. Those shapes, and how they interact, control how the machine works. Understanding that has been a Holy Grail of biomedical science.
It took billions of dollars, thousands of scientists, and a century of effort to determine the shape of 30% of the proteins in humans. Some scientists spent their entire career identifying the shape of just one important protein. Then, in July 2021, DeepMind, a company owned by Google, shocked the biomedical world: using A.I., it had determined the shape of nearly all the human proteins (98%). Overnight, we went from knowing the shape of about 30% of human proteins to knowing the shape of nearly all of them.
Even more remarkable, there are 280 million known non-human proteins in plants and in other animals. A century's work had revealed the structure of only 0.01% of them. DeepMind's A.I. program now has determined the structure of nearly all of those proteins — from 0.01% to nearly 100%! That's an astonishing breakthrough.
The ability of A.I. to predict the shape of each human protein, remarkable as it was, was only the first step. Skeptics doubted that A.I. could take the next step: figuring out how specific proteins interact with other proteins and with other molecules. Now DeepMind's scientists have proved the skeptics wrong, with evidence published in May 2024 in the prestigious journal Nature.
Most experts believe this A.I. advance will speed the development of many potent medicines and vaccines, and cures for some diseases. Time will tell if this forecast is too optimistic. But no one now doubts that A.I. has produced a breakthrough. And it could only have happened because our society invested in the sciences of information technology and biomedicine.
Image: © sanjeri/Getty Images
About the Author
Anthony L. Komaroff, MD, Editor in Chief, Harvard Health Letter; Editorial Advisory Board Member, Harvard Health Publishing
Disclaimer:
As a service to our readers, Harvard Health Publishing provides access to our library of archived content. Please note the date of last review or update on all articles.
No content on this site, regardless of date, should ever be used as a substitute for direct medical advice from your doctor or other qualified clinician.
