AI model decodes chemical signals in embryonic cell development

Researchers have developed a machine learning model capable of decoding the chemical signals that guide embryonic cell development. During embryonic growth, stem cells receive signals through multi-step sequences known as signaling pathways, which dictate whether a cell becomes specialized tissue, such as muscle, brain, or liver cells.

Previously, scientists believed the effects of these signaling pathways varied so significantly across different cell types that each would need to be mapped individually. However, a study led by Whitehead Institute member Pulin Li and graduate student Nicholas Hutchins has revealed that each signaling pathway leaves a unique “fingerprint” of gene activity.

Because these fingerprints remain consistent across different cell types for the same pathway, researchers can now reconstruct signaling histories across various cell types without the need for separate, painstaking mapping for every individual cell type. This discovery may help scientists understand how tissue development occurs and how these processes can malfunction in disease.