In 1889, French physician Francois-Gilbert Viault came down from a mountain in the Andes, drew blood from his arm and examined it under a microscope. Dr. Violt's red blood cell count, which carries oxygen, jumped by 42 percent. He discovered the mysterious powers of the human body. The idea is that when you need more of these important cells, you can make them on demand.
In the early 1900s, scientists theorized that hormones were the cause. They called the theoretical hormone erythropoietin, or “red maker” in Greek. Seventy years later, researchers discovered actual erythropoietin after filtering 670 gallons of urine.
And about 50 years later, Israeli biologists announced the discovery of a rare kidney cell that produces hormones when oxygen levels drop too low. It is called a norn cell, named after the Norse gods who are believed to control human destiny.
It took humans 134 years to discover yolk cells. Last summer, a computer in California discovered it on its own in just six weeks.
The discovery was made when Stanford researchers programmed a computer to teach itself biology. The computer ran an artificial intelligence program similar to ChatGPT, a popular bot that became proficient in language after training on billions of texts from the Internet. But Stanford researchers trained the computer on millions of real cells and raw data about their chemical and genetic makeup.
The researchers did not tell the computer what these measurements meant. They did not account for the fact that different types of cells have different biochemical profiles. For example, they did not define which cells receive light in our eyes or which cells make antibodies.
The computer itself processed the data to create a model of every cell based on their similarities to each other in a vast multidimensional space. By the time the machine was completed, they had learned a surprising amount. They were able to classify cells they had never seen before into one of over 1,000 types. One of them was the Norn cell.
“This is surprising because no one has ever told the model that Norn cells exist in the kidney,” said Jure Leskovec, a computer scientist at Stanford University who trained the computer.
The software is one of several new AI-based programs known as foundational models and is concerned with the fundamentals of biology. Models don't simply organize the information biologists gather. They are making discoveries about how genes work and how cells develop.
As the model expands with more laboratory data and computing power, scientists predict that more profound discoveries will begin. They may reveal secrets about cancer and other diseases. They may even be able to figure out how to change one type of cell into another type of cell.
“This is an important discovery in biology that biologists would not have otherwise made,” said Dr. Eric Topol, director of the Scripps Research Translational Institute. “One day you will see this.”
How far they will go is a matter of debate. While some skeptics think the model will hit a wall, more optimistic scientists believe the basic model will even address the biggest biological question: What distinguishes living things from non-living things.