In this study, we adopt the perspective of a naive researcher with the goal of learning a new signaling pathway, starting from differential phosphoproteomic data and a kinase-substrate network. Focusing on the well-characterized epidermal growth factor (EGF) response, we perform a meta-analysis of recent studies and complement those with three deep and time-resolved phosphoproteomic experiments. We also combine multiple kinase-substrate networks using protein language models and data from peptide-array screenings, generating a state-of-the-art resource. We use these datasets and networks to generate context-specific signaling subnetworks through various computational approaches, and compare the results to different sets of ground truth interactions.