The paper introduces Tahoe-100M, a groundbreaking giga-scale single-cell perturbation atlas containing over 100 million transcriptomic profiles. This extensive dataset, generated using the Mosaic platform by Tahoe Therapeutics, maps how 1,100 small-molecule perturbations affect cells across 50 cancer cell lines. The research highlights the dataset's utility in developing AI-driven models for understanding context-dependent gene function and cellular modeling, offering unprecedented insights into cell biology and accelerating cancer drug discovery. The paper further explains the methodology behind data generation and demonstrates how Tahoe-100M reveals cellular heterogeneity in drug response, including pathway-specific effects and cell cycle alterations.
Reference:
