Parse Biosciences and Graph Therapeutics announced a strategic collaboration to create one of the largest immune‑cell perturbation atlases to date. The partnership will combine Graph’s patient‑derived disease models and active‑learning perturbation selection with Parse’s GigLab single‑cell transcriptomics platform to profile hundreds of millions of immune cells across disease contexts. The resulting dataset is intended to power AI‑first drug discovery by mapping context‑specific immune responses to systematic perturbations, improving target selection and translational fidelity. For drug developers, the atlas aims to derisk immunology programs and reduce late‑stage clinical surprises by prioritizing perturbations that show consistent, disease‑relevant effects in human primary cells. Clarification: 'Perturbation atlases' systematically test how cells respond to drugs, cytokines, or genetic edits to reveal actionable biology and potential therapeutic targets.