Models of ligand discrimination in cross-wired signaling pathways

Cellular signaling pathways exhibit remarkable specificity, discrimination and sensitivity. Puzzlingly, many signaling pathways common to the immune system are highly cross-wired whereby extracellular ligand molecules act through shared receptors or overlapping pathways downstream. It remains unclear how cells can reliably transmit information in such conditions especially when the ligands are present in multiple combinations. I will present several theoretical models of accurate and specific input-output mapping in cross-wired pathways, and show how a ligand-receptor signaling pathway can be mapped onto a classifier of an artificial neural network type. I will present the results of the combined experimental and computational testing of some of the theoretical models on the example of a major class of signaling molecules of the immune system, Type I Interferons (IFN), where multiple IFNs act through the same receptor. These results have important implications for the signaling by a broad class of cross-wired signaling pathways and for the clinical use of signaling molecules in disease treatment.