Combinatorial cell signals encode cell fate patterns

Understanding how combinatorial cell signaling controls cellular decisions in the face of cross-talk remains a fundamental challenge in biology. Here, we present a generalizable framework integrating iterative immunofluorescence, information theory, and machine learning to relate combinatorial cell signaling to cell fate. We apply it to human 2D gastruloids, a model for human embryonic tissue patterning, in which we simultaneously measured spatially-resolved measurements of six cell signaling pathways and expression of up to seventeen cell fate marker genes at the single-cell level. We introduce Sig2Fate, a comprehensive "signal-to-fate" map constructed from this data that accurately predicts fate marker gene expression from signaling - including after changes in exogenous ligand concentration and pharmacological perturbation of cell signaling. Sig2Fate reveals how cell fate is encoded by the collective state of multiple signaling pathways and maps out pathway redundancies, providing a handle on explaining complex phenotypes that result from signaling perturbations in many biological systems. Taken together, Sig2Fate provides a scalable solution for mapping complex signaling interactions to reveal the regulatory logic of cellular decision making.