The Chan Zuckerberg Initiative (CZI), the philanthropic organization created in 2015 by Priscilla Chan and her husband Mark Zuckerberg, announced a bold new generative AI initiative today. The group is funding and building a high-end GPU cluster that will use AI to create predictive models of healthy and diseased cells; it hopes they’ll help researchers better understand the human body’s cells and cellular reactions. The group believes the collection of computers will help it achieve its incredibly lofty goal of helping to “cure, prevent, or manage all diseases by the end of this century.”
“Researchers are gathering more data than ever before about the trillions of cells within our bodies, and it’s too complex for our brains to grapple with,” Jeff MacGregor, CZI vice president of communications, wrote in an emailed statement to Engadget. He lists an example of imaging one cell at nanometer resolution, which would use the same amount of data as 83,000 photos on a smartphone. Sifting through the finer details of a cache of cellular models like that is where generative AI could play a role.
The system will use a cluster of over 1,000 GPUs to train AI large language models (LLMs) on human cells. “LLMs have done an impressive job at helping us understand protein structure, and we think they will be equally great at helping us understand more complex structures like cells,” MacGregor said. He expects the AI models to draw insights and conclusions beyond even the capabilities of a team of human experts. “But also, it’s about the speed at which they can do this. It would take that team of experts years to draw the types of insights rather than weeks that it will take for the models to do so.”
Chan lists other examples of how LLMs could tackle biomedicine’s problems. “AI models could predict how an immune cell responds to an infection, what happens at the cellular level when a child is born with a rare disease, or even how a patient’s body will respond to a new medication,” the co-founder and co-CEO said. “We hope that this collaborative effort will generate new insights about the fundamental characteristics of our cells.”
The group describes the GPU clusters as one of the first to power “openly available” models of human cells, suggesting the investment could yield dividends for under-funded researchers with bright ideas. Examples of data the models will train on include those integrated into the Chan Zuckerberg Cell by Gene tool (with its existing database of over 50 million cells), resources from CZ Science research institutes and publicly available datasets. CZI Head of Science Stephen Quake describes one of the project’s goals as creating a “virtual biology simulator.”
“AI is creating new opportunities in biomedicine, and building a high-performance computing cluster dedicated to life science research will accelerate progress on important scientific questions about how our cells work,” said Zuckerberg. “Developing digital models capable of predicting all cell types and cell states from the genome will help researchers better understand our cells and how they behave in health and disease.”