RESEARCH · RESEARCH · #2070
AI model IRIS decodes signaling-pathway 'fingerprints' in embryonic cells
MIT and Whitehead researchers developed IRIS, a neural‑network model that detects pathway‑specific gene‑expression 'fingerprints' to infer which signaling pathways cells experienced; trained on large human embryonic stem‑cell datasets and validated on single cells from mouse gastrulation, the open‑access study was published in Nature Methods on Sept. 8.
KEY POINTS
- MIT and Whitehead researchers developed IRIS, a neural‑network model that detects pathway‑specific gene‑expression 'fingerprints' to infer which signaling pathways cells experienced; trained on large human embryonic stem‑cell datasets and validated on single cells from mouse gastrulation, the open‑access study was published in Nature Methods on Sept.
- IRIS lets researchers reconstruct signaling histories across many cell types without mapping each pathway in each cell type, enabling scalable advances in developmental biology, stem‑cell engineering, and organoid production.
- AI model decodes the language cells use to communicate
WHY IT MATTERS
IRIS lets researchers reconstruct signaling histories across many cell types without mapping each pathway in each cell type, enabling scalable advances in developmental biology, stem‑cell engineering, and organoid production.