Physics / Math
Network motifs (Milo et al., Science 2002) are small subgraphs that appear significantly more often than in randomized null models. The feed-forward loop (FFL): A→B, A→C, B→C — appears ~10× over-represented in transcription networks and acts as a sign-sensitive delay/filter. The bi-fan motif (two inputs → two outputs) is common in signal transduction. Z-score = (N_observed − ⟨N_random⟩) / σ_random quantifies significance. Motifs define the "design principles" of biological networks.