University of Toronto researchers map protein network dynamics during cell division

An international team led by University of Toronto researchers has mapped the movement of proteins encoded by the yeast genome throughout the cell cycle. This is the first time all proteins of an organism have been tracked throughout the cell cycle, requiring a combination of deep learning and high-throughput microscopy.

The team applied two convolutional neural networks, or algorithms, called DeepLoc and CycleNet, to analyze millions of images of living yeast cells. The result is a comprehensive map that identifies the location of proteins and how they move within the cell and change in abundance during each stage of the cell cycle.

“We found that proteins whose intracellular concentrations regularly increase and decrease tend to be involved in regulating the cell cycle, while proteins with predictable movements in cells tend to help promote the biophysical implementation of the cell cycle,” said Athanasios Liziosthe study’s first author and a postdoctoral researcher at the University of Toronto’s Donnelly Center for Cellular and Biomolecular Research.

The research was recently published in the journal cell.

The cell cycle is understood as the stage in which cells progress to their eventual division into individual cells. It is this process that forms the basis for the proliferation of life and is ongoing in all living things.

At a molecular level, the cell cycle depends on the coordination of many proteins to guide cells from growth and DNA replication through cell division. Dysregulation of proteins can disrupt the cell cycle, and their disruption can lead to diseases such as cancer.

The researchers observed that about a quarter of the mapped yeast proteins followed regular patterns of appearing and disappearing or moving to specific regions of the cell. Most proteins follow these concentration or movement patterns, but not both.

“We identified about 400 proteins that have only periodic localization in the cell cycle, and about 800 proteins that have only periodic concentration,” Litsios said. “This means that the protein is regulated at multiple levels to ensure that the cell cycle proceeds as programmed.”

The research team used fluorescence microscopy to track approximately 4,000 proteins in yeast cell images to classify cell cycle stages and the location of proteins within 22 classified regions including the nucleus, cytoplasm and mitochondria. By automating phase and protein position identification using convolutional neural networks, cell cycle phase prediction was achieved with an accuracy of over 93%.

“We analyzed more than 20 million images of live yeast cells and used machine learning to assign them to different cell cycle stages,” said Brenda Andrews, the study’s principal investigator and University Professor of Molecular Genetics at the Donnelly Center and Temerty School of Medicine. “We then developed and applied a second computational pipeline to investigate how the protein’s localization and concentration change during the cell cycle. This study produced a unique dataset that provides a genome-wide picture of the molecular changes that occur during cell division. Scale view.

“Yeast cells are a great model for eukaryotic biology,” Lizios said. “We can do some things with yeast cells that we can’t do with other simpler or more complex organisms. We can use yeast cells to observe processes on a large scale, which makes it the perfect organism to study the cell cycle and hopefully better to better understand the human cell cycle.

This research was supported by the Canada Foundation for Innovation, the Canadian Institutes of Health Research, the National Institutes of Health and the Ontario Research Fund.