Stowers Institute, Helmholtz Munich, Technical University of Munich, and the University of Oxford unveiled RegVelo, an AI model designed to infer gene-regulatory networks driving single-cell fate transitions over time. Published in Cell, the framework connects how cells move through developmental trajectories with the regulatory instructions that govern those transitions. In zebrafish neural crest development, RegVelo identified an early driver of pigment cell formation (tfec) and suggested a previously unknown regulator of pigment fate (elf1). The predictions were supported with CRISPR/Cas9 knockouts and single-cell Perturb-seq, showing the model can generate testable hypotheses that hold up in living systems. The teams also highlighted potential use beyond development, including modeling tumor trajectories where fate changes and gene programs evolve under selection pressure.