Teaching a computer program to track cells


Scientists at the Gladstone Institutes have developed such an approach, which uses “neural networks” – artificial intelligence programs capable of detecting patterns – to analyze the location of hundreds of cells growing together in a colony. When they applied the technique to a group of stem cells, the program found that a small number of cells act as “leaders,” able to direct the movements of their neighbors.

“This technique gives us a much more complete view of how cells behave, how they cooperate and how they come together in physical space to form complex organs,” says Gladstone Senior Investigator. Todd C. McDevitt, PhD, main author of a new article published in the journal Stem cell reports.

Clusters of stem cells have the ability to form any tissue in the human body when exposed to the right mix of signaling molecules. But researchers have a poor understanding of how these cells form patterns in space to eventually give rise to complex three-dimensional organs.

Traditionally, to study how cells move in space over time, cell biologists tag cells with fluorescent molecules that make them easy to follow. Then they observe these cells under a microscope to see how they divide and migrate. However, a human observer can only follow a small handful of cells at a time before it becomes too difficult to distinguish different cells and track their movements. This means that scientists often have to extrapolate how an entire colony moves based on the movements of a few of its members.

In the new paper, McDevitt’s group trained three different neural networks to track the movements of individual cells within colonies of thousands of cells. Each network had its own strengths and weaknesses, and individually none of them outperformed one person. But combined, the three neural networks were slightly more efficient at tracking cells: They were able to find 94% of all cells in sequential frames, meaning they could track their movements over time. Humans could only follow 90% of all cells between images; a scientist trying to track cell movements could only figure out where nine out of ten cells were moving. Additionally, the combined neural networks were 500 times faster than a person, averaging 0.35 seconds per image to identify all cells, while a human averaged about 3 minutes per image.

When the researchers used the arrays to study new stem cell colonies, they were surprised to see much more action than previous cell-tracking techniques had identified. While the colonies appeared fairly static to the naked eye, neural networks showed that almost all of the cells were in motion and much of the movement appeared random.

“When we walked in, we weren’t really expecting there to be so much cell movement, so we had to come up with new approaches to understanding the apparent chaos of cells,” says David Joy, Gladstone graduate student, lead author. of the new article. .

Cells closest to the edges of each colony moved the most, McDevitt’s group found. And cells tended to start and stop more than the researchers would have guessed – on average, each cell moved for about 15 minutes, followed by a 10-minute rest and stillness period. before the start of another active phase.

The researchers then showed how changing the conditions in the cells’ environment, by exposing the cells to different nutrients or drugs, can change the way cells move. They also used neural networks to track stem cell colonies for 24 hours as they began to form the multiple layers of different cell types that appear in an early embryo. The team found that cells have a wide range of movement patterns.

“Some cells move very persistently in one direction, while others move endlessly but never stray from their starting point,” explains Ashley libby, PhD, a former graduate student of McDevitt’s lab who helped direct the work. The diversity surprised the team; they expected most cells to follow similar movement patterns, she says.

In addition, some cells acted as “leaders” while others behaved more like “followers”, according to the researchers. The movement of a small number of cells spread to their neighbors, ultimately changing the dynamics of the entire colony. This is a pattern that would not have been evident if only a few cells had been followed over time by a human observer.

The new findings are just a small sampling of the types of observations that will be possible when artificial intelligence approaches are applied to cell tracking, McDevitt says. And the knowledge that emerges from these future experiments will be useful to researchers trying to coax cells into assembling into complex organoids and organs, both for research and therapeutic purposes.

“If I wanted to create a new human heart now, I know what types of cells are needed and I know how to grow them independently in dishes,” says McDevitt, who is also a professor of bioengineering and therapeutic sciences at UC. San. Francois. “But we really don’t know how to get these cells to come together to form something as complex as a heart. To do this, we need more information about how cells work cooperatively. to organize. “

In a next step, McDevitt’s team is planning future studies using neural networks to analyze the movements within cultures of stem cells with genetic mutations, in order to show the effect of different genes on cell organization.

About the study

The article “Deep neural network monitoring of human pluripotent stem cells reveals intrinsic behaviors directing morphogenesis” was published by the journal Stem cell reports on May 11, 2021.

The work was supported by the California Institute of Regenerative Medicine (LA1_C14-08015) and the National Science Foundation (CBET 0939511).

About Gladstone Institutes

To ensure that our work does the greatest good, the Gladstone Institutes focus on conditions with profound medical, economic and social impact – unresolved illnesses. Gladstone is an independent, non-profit life science research organization that uses visionary science and technology to beat disease. He has an academic affiliation with the University of California, San Francisco.

Media contact: Julie langelier | Assistant Director, Communications | [email protected] | 415,734,500

Gladstone Institutes

Related links



Gordon K. Morehouse