AI model IRIS decodes the signals cells use to communicate during development
Original titleAI model decodes the language cells use to communicate
Researchers led by Pulin Li and Nicholas Hutchins at MIT and Whitehead Institute built IRIS, a neural network that identifies which of six signaling pathways a cell received during embryonic development.
The model was trained on human embryonic stem cell data, predicted signaling in mouse gastrulation cells, and correctly predicted that activating a specific pathway would promote lung development, which mouse experiments confirmed.
The findings were published in Nature Methods on September 8.
The article explains how a machine learning model reads signaling fingerprints to reconstruct cell histories, a method that could make stem cell and organoid work less experiment-heavy.
Source: MIT News · AI · news.mit.eduPublished · added here