Chan Zuckerberg Initiative Builds GPU Supercomputer and Cell Atlas for AI Virtual Cell Models
Start here
The Chan Zuckerberg Initiative is developing an AI-driven virtual cell model to simulate how human cells function, interact and form organs. CZI science head Steve Quake describes genomes and protein structures as largely solved problems, with cell behavior the remaining frontier. To train such models, CZI is building one of the largest nonprofit GPU supercomputers dedicated to basic research and has assembled expert-annotated data on more than 30 million human cells, drawn from Tabula Sapiens and global tissue atlases and integrated through CELLxGENE. Quake expects machine learning to eventually enable in silico experiments, such as predicting the whole-body cellular effects of a single mutation, and to support engineered cells that detect early tumours or neurodegenerative plaques. The supercomputer build-out and model training are ongoing, with no released virtual cell model yet.
Written by a model from the reports · updated 3 hours ago
Coverage timeline
Follow the coverage to see the different angles.
- youtube.comCZI Builds GPU Supercomputing and Cell Atlas to Train AI Virtual Cell Models
CZI science head Steve Quake says genomes and protein structures are largely solved problems, but how cells work, interact and form organs remains unknown. CZI is building one of the largest nonprofit GPU supercomputers for basic research and has assembled expert-annotated data on more than 30 million human cells, drawn from Tabula Sapiens and global tissue atlases and integrated through CELLxGENE. He expects machine learning to enable in silico experiments, such as predicting whole-body cellular effects of single mutations and using engineered cells to detect early tumours or neurodegenerative plaques.
Heat trend for this story
Not enough continuous observation yet to draw a trend.